/[pcre]/code/trunk/pcretest.c
ViewVC logotype

Diff of /code/trunk/pcretest.c

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 355 by ph10, Mon Jul 7 17:45:23 2008 UTC revision 839 by zherczeg, Fri Dec 30 13:22:28 2011 UTC
# Line 4  Line 4 
4    
5  /* This program was hacked up as a tester for PCRE. I really should have  /* This program was hacked up as a tester for PCRE. I really should have
6  written it more tidily in the first place. Will I ever learn? It has grown and  written it more tidily in the first place. Will I ever learn? It has grown and
7  been extended and consequently is now rather, er, *very* untidy in places.  been extended and consequently is now rather, er, *very* untidy in places. The
8    addition of 16-bit support has made it even worse. :-(
9    
10  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
11  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 35  POSSIBILITY OF SUCH DAMAGE. Line 36  POSSIBILITY OF SUCH DAMAGE.
36  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
37  */  */
38    
39    /* This program now supports the testing of both the 8-bit and 16-bit PCRE
40    libraries in a single program. This is different from the modules such as
41    pcre_compile.c in the library itself, which are compiled separately for each
42    mode. If both modes are enabled, for example, pcre_compile.c is compiled twice
43    (the second time with COMPILE_PCRE16 defined). By contrast, pcretest.c is
44    compiled only once. Therefore, it must not make use of any of the macros from
45    pcre_internal.h that depend on COMPILE_PCRE8 or COMPILE_PCRE16. It does,
46    however, make use of SUPPORT_PCRE8 and SUPPORT_PCRE16 to ensure that it calls
47    only supported library functions. */
48    
49    
50  #ifdef HAVE_CONFIG_H  #ifdef HAVE_CONFIG_H
51  #include "config.h"  #include "config.h"
# Line 71  input mode under Windows. */ Line 82  input mode under Windows. */
82  #define INPUT_MODE   "r"  #define INPUT_MODE   "r"
83  #define OUTPUT_MODE  "wb"  #define OUTPUT_MODE  "wb"
84    
85  #define isatty _isatty         /* This is what Windows calls them, I'm told */  #ifndef isatty
86    #define isatty _isatty         /* This is what Windows calls them, I'm told, */
87    #endif                         /* though in some environments they seem to   */
88                                   /* be already defined, hence the #ifndefs.    */
89    #ifndef fileno
90  #define fileno _fileno  #define fileno _fileno
91    #endif
92    
93    /* A user sent this fix for Borland Builder 5 under Windows. */
94    
95    #ifdef __BORLANDC__
96    #define _setmode(handle, mode) setmode(handle, mode)
97    #endif
98    
99    /* Not Windows */
100    
101  #else  #else
102  #include <sys/time.h>          /* These two includes are needed */  #include <sys/time.h>          /* These two includes are needed */
# Line 92  here before pcre_internal.h so that the Line 116  here before pcre_internal.h so that the
116  appropriately for an application, not for building PCRE. */  appropriately for an application, not for building PCRE. */
117    
118  #include "pcre.h"  #include "pcre.h"
119    
120    #if defined SUPPORT_PCRE16 && !defined SUPPORT_PCRE8
121    /* Configure internal macros to 16 bit mode. */
122    #define COMPILE_PCRE16
123    #endif
124    
125  #include "pcre_internal.h"  #include "pcre_internal.h"
126    
127    /* The pcre_printint() function, which prints the internal form of a compiled
128    regex, is held in a separate file so that (a) it can be compiled in either
129    8-bit or 16-bit mode, and (b) it can be #included directly in pcre_compile.c
130    when that is compiled in debug mode. */
131    
132    #ifdef SUPPORT_PCRE8
133    void pcre_printint(pcre *external_re, FILE *f, BOOL print_lengths);
134    #endif
135    #ifdef SUPPORT_PCRE16
136    void pcre16_printint(pcre *external_re, FILE *f, BOOL print_lengths);
137    #endif
138    
139  /* We need access to some of the data tables that PCRE uses. So as not to have  /* We need access to some of the data tables that PCRE uses. So as not to have
140  to keep two copies, we include the source file here, changing the names of the  to keep two copies, we include the source file here, changing the names of the
141  external symbols to prevent clashes. */  external symbols to prevent clashes. */
142    
143  #define _pcre_ucp_gentype      ucp_gentype  #define PCRE_INCLUDED
144  #define _pcre_utf8_table1      utf8_table1  #undef PRIV
145  #define _pcre_utf8_table1_size utf8_table1_size  #define PRIV(name) name
 #define _pcre_utf8_table2      utf8_table2  
 #define _pcre_utf8_table3      utf8_table3  
 #define _pcre_utf8_table4      utf8_table4  
 #define _pcre_utt              utt  
 #define _pcre_utt_size         utt_size  
 #define _pcre_utt_names        utt_names  
 #define _pcre_OP_lengths       OP_lengths  
146    
147  #include "pcre_tables.c"  #include "pcre_tables.c"
148    
149  /* We also need the pcre_printint() function for printing out compiled  /* The definition of the macro PRINTABLE, which determines whether to print an
 patterns. This function is in a separate file so that it can be included in  
 pcre_compile.c when that module is compiled with debugging enabled.  
   
 The definition of the macro PRINTABLE, which determines whether to print an  
150  output character as-is or as a hex value when showing compiled patterns, is  output character as-is or as a hex value when showing compiled patterns, is
151  contained in this file. We uses it here also, in cases when the locale has not  the same as in the printint.src file. We uses it here in cases when the locale
152  been explicitly changed, so as to get consistent output from systems that  has not been explicitly changed, so as to get consistent output from systems
153  differ in their output from isprint() even in the "C" locale. */  that differ in their output from isprint() even in the "C" locale. */
154    
155  #include "pcre_printint.src"  #ifdef EBCDIC
156    #define PRINTABLE(c) ((c) >= 64 && (c) < 255)
157    #else
158    #define PRINTABLE(c) ((c) >= 32 && (c) < 127)
159    #endif
160    
161  #define PRINTHEX(c) (locale_set? isprint(c) : PRINTABLE(c))  #define PRINTOK(c) (locale_set? isprint(c) : PRINTABLE(c))
162    
163    /* Posix support is disabled in 16 bit only mode. */
164    #if defined SUPPORT_PCRE16 && !defined SUPPORT_PCRE8 && !defined NOPOSIX
165    #define NOPOSIX
166    #endif
167    
168  /* It is possible to compile this test program without including support for  /* It is possible to compile this test program without including support for
169  testing the POSIX interface, though this is not available via the standard  testing the POSIX interface, though this is not available via the standard
# Line 134  Makefile. */ Line 173  Makefile. */
173  #include "pcreposix.h"  #include "pcreposix.h"
174  #endif  #endif
175    
176  /* It is also possible, for the benefit of the version currently imported into  /* It is also possible, originally for the benefit of a version that was
177  Exim, to build pcretest without support for UTF8 (define NOUTF8), without the  imported into Exim, to build pcretest without support for UTF8 or UTF16 (define
178  interface to the DFA matcher (NODFA), and without the doublecheck of the old  NOUTF), without the interface to the DFA matcher (NODFA). In fact, we
179  "info" function (define NOINFOCHECK). In fact, we automatically cut out the  automatically cut out the UTF support if PCRE is built without it. */
180  UTF8 support if PCRE is built without it. */  
181    #ifndef SUPPORT_UTF
182  #ifndef SUPPORT_UTF8  #ifndef NOUTF
183  #ifndef NOUTF8  #define NOUTF
 #define NOUTF8  
184  #endif  #endif
185  #endif  #endif
186    
187    /* To make the code a bit tidier for 8-bit and 16-bit support, we define macros
188    for all the pcre[16]_xxx functions (except pcre16_fullinfo, which is called
189    only from one place and is handled differently). I couldn't dream up any way of
190    using a single macro to do this in a generic way, because of the many different
191    argument requirements. We know that at least one of SUPPORT_PCRE8 and
192    SUPPORT_PCRE16 must be set. First define macros for each individual mode; then
193    use these in the definitions of generic macros.
194    
195    **** Special note about the PCHARSxxx macros: the address of the string to be
196    printed is always given as two arguments: a base address followed by an offset.
197    The base address is cast to the correct data size for 8 or 16 bit data; the
198    offset is in units of this size. If the string were given as base+offset in one
199    argument, the casting might be incorrectly applied. */
200    
201    #ifdef SUPPORT_PCRE8
202    
203    #define PCHARS8(lv, p, offset, len, f) \
204      lv = pchars((pcre_uint8 *)(p) + offset, len, f)
205    
206    #define PCHARSV8(p, offset, len, f) \
207      (void)pchars((pcre_uint8 *)(p) + offset, len, f)
208    
209    #define READ_CAPTURE_NAME8(p, cn8, cn16, re) \
210      p = read_capture_name8(p, cn8, re)
211    
212    #define SET_PCRE_CALLOUT8(callout) \
213      pcre_callout = callout
214    
215    #define STRLEN8(p) ((int)strlen((char *)p))
216    
217    
218    #define PCRE_COMPILE8(re, pat, options, error, erroffset, tables) \
219      re = pcre_compile((char *)pat, options, error, erroffset, tables)
220    
221    #define PCRE_COPY_NAMED_SUBSTRING8(rc, re, bptr, offsets, count, \
222        namesptr, cbuffer, size) \
223      rc = pcre_copy_named_substring(re, (char *)bptr, offsets, count, \
224        (char *)namesptr, cbuffer, size)
225    
226    #define PCRE_COPY_SUBSTRING8(rc, bptr, offsets, count, i, cbuffer, size) \
227      rc = pcre_copy_substring((char *)bptr, offsets, count, i, cbuffer, size)
228    
229    #define PCRE_DFA_EXEC8(count, re, extra, bptr, len, start_offset, options, \
230        offsets, size_offsets, workspace, size_workspace) \
231      count = pcre_dfa_exec(re, extra, (char *)bptr, len, start_offset, options, \
232        offsets, size_offsets, workspace, size_workspace)
233    
234    #define PCRE_EXEC8(count, re, extra, bptr, len, start_offset, options, \
235        offsets, size_offsets) \
236      count = pcre_exec(re, extra, (char *)bptr, len, start_offset, options, \
237        offsets, size_offsets)
238    
239    #define PCRE_FREE_STUDY8(extra) \
240      pcre_free_study(extra)
241    
242    #define PCRE_FREE_SUBSTRING8(substring) \
243      pcre_free_substring(substring)
244    
245    #define PCRE_FREE_SUBSTRING_LIST8(listptr) \
246      pcre_free_substring_list(listptr)
247    
248    #define PCRE_GET_NAMED_SUBSTRING8(rc, re, bptr, offsets, count, \
249        getnamesptr, subsptr) \
250      rc = pcre_get_named_substring(re, (char *)bptr, offsets, count, \
251        (char *)getnamesptr, subsptr)
252    
253    #define PCRE_GET_STRINGNUMBER8(n, rc, ptr) \
254      n = pcre_get_stringnumber(re, (char *)ptr)
255    
256    #define PCRE_GET_SUBSTRING8(rc, bptr, offsets, count, i, subsptr) \
257      rc = pcre_get_substring((char *)bptr, offsets, count, i, subsptr)
258    
259    #define PCRE_GET_SUBSTRING_LIST8(rc, bptr, offsets, count, listptr) \
260      rc = pcre_get_substring_list((const char *)bptr, offsets, count, listptr)
261    
262    #define PCRE_PATTERN_TO_HOST_BYTE_ORDER8(rc, re, extra, tables) \
263      rc = pcre_pattern_to_host_byte_order(re, extra, tables)
264    
265    #define PCRE_PRINTINT8(re, outfile, debug_lengths) \
266      pcre_printint(re, outfile, debug_lengths)
267    
268    #define PCRE_STUDY8(extra, re, options, error) \
269      extra = pcre_study(re, options, error)
270    
271    #endif /* SUPPORT_PCRE8 */
272    
273    /* -----------------------------------------------------------*/
274    
275    #ifdef SUPPORT_PCRE16
276    
277    #define PCHARS16(lv, p, offset, len, f) \
278      lv = pchars16((PCRE_SPTR16)(p) + offset, len, f)
279    
280    #define PCHARSV16(p, offset, len, f) \
281      (void)pchars16((PCRE_SPTR16)(p) + offset, len, f)
282    
283    #define READ_CAPTURE_NAME16(p, cn8, cn16, re) \
284      p = read_capture_name16(p, cn16, re)
285    
286    #define STRLEN16(p) ((int)strlen16((PCRE_SPTR16)p))
287    
288    #define SET_PCRE_CALLOUT16(callout) \
289      pcre16_callout = callout
290    
291    
292    #define PCRE_COMPILE16(re, pat, options, error, erroffset, tables) \
293      re = pcre16_compile((PCRE_SPTR16)pat, options, error, erroffset, tables)
294    
295    #define PCRE_COPY_NAMED_SUBSTRING16(rc, re, bptr, offsets, count, \
296        namesptr, cbuffer, size) \
297      rc = pcre16_copy_named_substring(re, (PCRE_SPTR16)bptr, offsets, count, \
298        (PCRE_SPTR16)namesptr, (PCRE_SCHAR16 *)cbuffer, size/2)
299    
300    #define PCRE_COPY_SUBSTRING16(rc, bptr, offsets, count, i, cbuffer, size) \
301      rc = pcre16_copy_substring((PCRE_SPTR16)bptr, offsets, count, i, \
302        (PCRE_SCHAR16 *)cbuffer, size/2)
303    
304    #define PCRE_DFA_EXEC16(count, re, extra, bptr, len, start_offset, options, \
305        offsets, size_offsets, workspace, size_workspace) \
306      count = pcre16_dfa_exec(re, extra, (PCRE_SPTR16)bptr, len, start_offset, \
307        options, offsets, size_offsets, workspace, size_workspace)
308    
309    #define PCRE_EXEC16(count, re, extra, bptr, len, start_offset, options, \
310        offsets, size_offsets) \
311      count = pcre16_exec(re, extra, (PCRE_SPTR16)bptr, len, start_offset, \
312        options, offsets, size_offsets)
313    
314    #define PCRE_FREE_STUDY16(extra) \
315      pcre16_free_study(extra)
316    
317    #define PCRE_FREE_SUBSTRING16(substring) \
318      pcre16_free_substring((PCRE_SPTR16)substring)
319    
320    #define PCRE_FREE_SUBSTRING_LIST16(listptr) \
321      pcre16_free_substring_list((PCRE_SPTR16 *)listptr)
322    
323    #define PCRE_GET_NAMED_SUBSTRING16(rc, re, bptr, offsets, count, \
324        getnamesptr, subsptr) \
325      rc = pcre16_get_named_substring(re, (PCRE_SPTR16)bptr, offsets, count, \
326        (PCRE_SPTR16)getnamesptr, (PCRE_SPTR16 *)(void*)subsptr)
327    
328    #define PCRE_GET_STRINGNUMBER16(n, rc, ptr) \
329      n = pcre16_get_stringnumber(re, (PCRE_SPTR16)ptr)
330    
331    #define PCRE_GET_SUBSTRING16(rc, bptr, offsets, count, i, subsptr) \
332      rc = pcre16_get_substring((PCRE_SPTR16)bptr, offsets, count, i, \
333        (PCRE_SPTR16 *)(void*)subsptr)
334    
335    #define PCRE_GET_SUBSTRING_LIST16(rc, bptr, offsets, count, listptr) \
336      rc = pcre16_get_substring_list((PCRE_SPTR16)bptr, offsets, count, \
337        (PCRE_SPTR16 **)(void*)listptr)
338    
339    #define PCRE_PATTERN_TO_HOST_BYTE_ORDER16(rc, re, extra, tables) \
340      rc = pcre16_pattern_to_host_byte_order(re, extra, tables)
341    
342    #define PCRE_PRINTINT16(re, outfile, debug_lengths) \
343      pcre16_printint(re, outfile, debug_lengths)
344    
345    #define PCRE_STUDY16(extra, re, options, error) \
346      extra = pcre16_study(re, options, error)
347    
348    #endif /* SUPPORT_PCRE16 */
349    
350    
351    /* ----- Both modes are supported; a runtime test is needed, except for
352    pcre_config(), and the JIT stack functions, when it doesn't matter which
353    version is called. ----- */
354    
355    #if defined SUPPORT_PCRE8 && defined SUPPORT_PCRE16
356    
357    #define CHAR_SIZE (use_pcre16? 2:1)
358    
359    #define PCHARS(lv, p, offset, len, f) \
360      if (use_pcre16) \
361        PCHARS16(lv, p, offset, len, f); \
362      else \
363        PCHARS8(lv, p, offset, len, f)
364    
365    #define PCHARSV(p, offset, len, f) \
366      if (use_pcre16) \
367        PCHARSV16(p, offset, len, f); \
368      else \
369        PCHARSV8(p, offset, len, f)
370    
371    #define READ_CAPTURE_NAME(p, cn8, cn16, re) \
372      if (use_pcre16) \
373        READ_CAPTURE_NAME16(p, cn8, cn16, re); \
374      else \
375        READ_CAPTURE_NAME8(p, cn8, cn16, re)
376    
377    #define SET_PCRE_CALLOUT(callout) \
378      if (use_pcre16) \
379        SET_PCRE_CALLOUT16(callout); \
380      else \
381        SET_PCRE_CALLOUT8(callout)
382    
383    #define STRLEN(p) (use_pcre16? STRLEN16(p) : STRLEN8(p))
384    
385    #define PCRE_ASSIGN_JIT_STACK pcre_assign_jit_stack
386    
387    #define PCRE_COMPILE(re, pat, options, error, erroffset, tables) \
388      if (use_pcre16) \
389        PCRE_COMPILE16(re, pat, options, error, erroffset, tables); \
390      else \
391        PCRE_COMPILE8(re, pat, options, error, erroffset, tables)
392    
393    #define PCRE_CONFIG pcre_config
394    
395    #define PCRE_COPY_NAMED_SUBSTRING(rc, re, bptr, offsets, count, \
396        namesptr, cbuffer, size) \
397      if (use_pcre16) \
398        PCRE_COPY_NAMED_SUBSTRING16(rc, re, bptr, offsets, count, \
399          namesptr, cbuffer, size); \
400      else \
401        PCRE_COPY_NAMED_SUBSTRING8(rc, re, bptr, offsets, count, \
402          namesptr, cbuffer, size)
403    
404    #define PCRE_COPY_SUBSTRING(rc, bptr, offsets, count, i, cbuffer, size) \
405      if (use_pcre16) \
406        PCRE_COPY_SUBSTRING16(rc, bptr, offsets, count, i, cbuffer, size); \
407      else \
408        PCRE_COPY_SUBSTRING8(rc, bptr, offsets, count, i, cbuffer, size)
409    
410    #define PCRE_DFA_EXEC(count, re, extra, bptr, len, start_offset, options, \
411        offsets, size_offsets, workspace, size_workspace) \
412      if (use_pcre16) \
413        PCRE_DFA_EXEC16(count, re, extra, bptr, len, start_offset, options, \
414          offsets, size_offsets, workspace, size_workspace); \
415      else \
416        PCRE_DFA_EXEC8(count, re, extra, bptr, len, start_offset, options, \
417          offsets, size_offsets, workspace, size_workspace)
418    
419    #define PCRE_EXEC(count, re, extra, bptr, len, start_offset, options, \
420        offsets, size_offsets) \
421      if (use_pcre16) \
422        PCRE_EXEC16(count, re, extra, bptr, len, start_offset, options, \
423          offsets, size_offsets); \
424      else \
425        PCRE_EXEC8(count, re, extra, bptr, len, start_offset, options, \
426          offsets, size_offsets)
427    
428    #define PCRE_FREE_STUDY(extra) \
429      if (use_pcre16) \
430        PCRE_FREE_STUDY16(extra); \
431      else \
432        PCRE_FREE_STUDY8(extra)
433    
434    #define PCRE_FREE_SUBSTRING(substring) \
435      if (use_pcre16) \
436        PCRE_FREE_SUBSTRING16(substring); \
437      else \
438        PCRE_FREE_SUBSTRING8(substring)
439    
440    #define PCRE_FREE_SUBSTRING_LIST(listptr) \
441      if (use_pcre16) \
442        PCRE_FREE_SUBSTRING_LIST16(listptr); \
443      else \
444        PCRE_FREE_SUBSTRING_LIST8(listptr)
445    
446    #define PCRE_GET_NAMED_SUBSTRING(rc, re, bptr, offsets, count, \
447        getnamesptr, subsptr) \
448      if (use_pcre16) \
449        PCRE_GET_NAMED_SUBSTRING16(rc, re, bptr, offsets, count, \
450          getnamesptr, subsptr); \
451      else \
452        PCRE_GET_NAMED_SUBSTRING8(rc, re, bptr, offsets, count, \
453          getnamesptr, subsptr)
454    
455    #define PCRE_GET_STRINGNUMBER(n, rc, ptr) \
456      if (use_pcre16) \
457        PCRE_GET_STRINGNUMBER16(n, rc, ptr); \
458      else \
459        PCRE_GET_STRINGNUMBER8(n, rc, ptr)
460    
461    #define PCRE_GET_SUBSTRING(rc, bptr, use_offsets, count, i, subsptr) \
462      if (use_pcre16) \
463        PCRE_GET_SUBSTRING16(rc, bptr, use_offsets, count, i, subsptr); \
464      else \
465        PCRE_GET_SUBSTRING8(rc, bptr, use_offsets, count, i, subsptr)
466    
467    #define PCRE_GET_SUBSTRING_LIST(rc, bptr, offsets, count, listptr) \
468      if (use_pcre16) \
469        PCRE_GET_SUBSTRING_LIST16(rc, bptr, offsets, count, listptr); \
470      else \
471        PCRE_GET_SUBSTRING_LIST8(rc, bptr, offsets, count, listptr)
472    
473    #define PCRE_JIT_STACK_ALLOC pcre_jit_stack_alloc
474    #define PCRE_JIT_STACK_FREE pcre_jit_stack_free
475    
476    #define PCRE_MAKETABLES \
477      (use_pcre16? pcre16_maketables() : pcre_maketables())
478    
479    #define PCRE_PATTERN_TO_HOST_BYTE_ORDER(rc, re, extra, tables) \
480      if (use_pcre16) \
481        PCRE_PATTERN_TO_HOST_BYTE_ORDER16(rc, re, extra, tables); \
482      else \
483        PCRE_PATTERN_TO_HOST_BYTE_ORDER8(rc, re, extra, tables)
484    
485    #define PCRE_PRINTINT(re, outfile, debug_lengths) \
486      if (use_pcre16) \
487        PCRE_PRINTINT16(re, outfile, debug_lengths); \
488      else \
489        PCRE_PRINTINT8(re, outfile, debug_lengths)
490    
491    #define PCRE_STUDY(extra, re, options, error) \
492      if (use_pcre16) \
493        PCRE_STUDY16(extra, re, options, error); \
494      else \
495        PCRE_STUDY8(extra, re, options, error)
496    
497    /* ----- Only 8-bit mode is supported ----- */
498    
499    #elif defined SUPPORT_PCRE8
500    #define CHAR_SIZE                 1
501    #define PCHARS                    PCHARS8
502    #define PCHARSV                   PCHARSV8
503    #define READ_CAPTURE_NAME         READ_CAPTURE_NAME8
504    #define SET_PCRE_CALLOUT          SET_PCRE_CALLOUT8
505    #define STRLEN                    STRLEN8
506    #define PCRE_ASSIGN_JIT_STACK     pcre_assign_jit_stack
507    #define PCRE_COMPILE              PCRE_COMPILE8
508    #define PCRE_CONFIG               pcre_config
509    #define PCRE_COPY_NAMED_SUBSTRING PCRE_COPY_NAMED_SUBSTRING8
510    #define PCRE_COPY_SUBSTRING       PCRE_COPY_SUBSTRING8
511    #define PCRE_DFA_EXEC             PCRE_DFA_EXEC8
512    #define PCRE_EXEC                 PCRE_EXEC8
513    #define PCRE_FREE_STUDY           PCRE_FREE_STUDY8
514    #define PCRE_FREE_SUBSTRING       PCRE_FREE_SUBSTRING8
515    #define PCRE_FREE_SUBSTRING_LIST  PCRE_FREE_SUBSTRING_LIST8
516    #define PCRE_GET_NAMED_SUBSTRING  PCRE_GET_NAMED_SUBSTRING8
517    #define PCRE_GET_STRINGNUMBER     PCRE_GET_STRINGNUMBER8
518    #define PCRE_GET_SUBSTRING        PCRE_GET_SUBSTRING8
519    #define PCRE_GET_SUBSTRING_LIST   PCRE_GET_SUBSTRING_LIST8
520    #define PCRE_JIT_STACK_ALLOC      pcre_jit_stack_alloc
521    #define PCRE_JIT_STACK_FREE       pcre_jit_stack_free
522    #define PCRE_MAKETABLES           pcre_maketables()
523    #define PCRE_PATTERN_TO_HOST_BYTE_ORDER PCRE_PATTERN_TO_HOST_BYTE_ORDER8
524    #define PCRE_PRINTINT             PCRE_PRINTINT8
525    #define PCRE_STUDY                PCRE_STUDY8
526    
527    /* ----- Only 16-bit mode is supported ----- */
528    
529    #else
530    #define CHAR_SIZE                 2
531    #define PCHARS                    PCHARS16
532    #define PCHARSV                   PCHARSV16
533    #define READ_CAPTURE_NAME         READ_CAPTURE_NAME16
534    #define SET_PCRE_CALLOUT          SET_PCRE_CALLOUT16
535    #define STRLEN                    STRLEN16
536    #define PCRE_ASSIGN_JIT_STACK     pcre16_assign_jit_stack
537    #define PCRE_COMPILE              PCRE_COMPILE16
538    #define PCRE_CONFIG               pcre16_config
539    #define PCRE_COPY_NAMED_SUBSTRING PCRE_COPY_NAMED_SUBSTRING16
540    #define PCRE_COPY_SUBSTRING       PCRE_COPY_SUBSTRING16
541    #define PCRE_DFA_EXEC             PCRE_DFA_EXEC16
542    #define PCRE_EXEC                 PCRE_EXEC16
543    #define PCRE_FREE_STUDY           PCRE_FREE_STUDY16
544    #define PCRE_FREE_SUBSTRING       PCRE_FREE_SUBSTRING16
545    #define PCRE_FREE_SUBSTRING_LIST  PCRE_FREE_SUBSTRING_LIST16
546    #define PCRE_GET_NAMED_SUBSTRING  PCRE_GET_NAMED_SUBSTRING16
547    #define PCRE_GET_STRINGNUMBER     PCRE_GET_STRINGNUMBER16
548    #define PCRE_GET_SUBSTRING        PCRE_GET_SUBSTRING16
549    #define PCRE_GET_SUBSTRING_LIST   PCRE_GET_SUBSTRING_LIST16
550    #define PCRE_JIT_STACK_ALLOC      pcre16_jit_stack_alloc
551    #define PCRE_JIT_STACK_FREE       pcre16_jit_stack_free
552    #define PCRE_MAKETABLES           pcre16_maketables()
553    #define PCRE_PATTERN_TO_HOST_BYTE_ORDER PCRE_PATTERN_TO_HOST_BYTE_ORDER16
554    #define PCRE_PRINTINT             PCRE_PRINTINT16
555    #define PCRE_STUDY                PCRE_STUDY16
556    #endif
557    
558    /* ----- End of mode-specific function call macros ----- */
559    
560    
561  /* Other parameters */  /* Other parameters */
562    
# Line 173  static int debug_lengths; Line 584  static int debug_lengths;
584  static int first_callout;  static int first_callout;
585  static int locale_set = 0;  static int locale_set = 0;
586  static int show_malloc;  static int show_malloc;
587  static int use_utf8;  static int use_utf;
588  static size_t gotten_store;  static size_t gotten_store;
589    static size_t first_gotten_store = 0;
590    static const unsigned char *last_callout_mark = NULL;
591    
592  /* The buffers grow automatically if very long input lines are encountered. */  /* The buffers grow automatically if very long input lines are encountered. */
593    
594  static int buffer_size = 50000;  static int buffer_size = 50000;
595  static uschar *buffer = NULL;  static pcre_uint8 *buffer = NULL;
596  static uschar *dbuffer = NULL;  static pcre_uint8 *dbuffer = NULL;
597  static uschar *pbuffer = NULL;  static pcre_uint8 *pbuffer = NULL;
598    
599    /* Another buffer is needed translation to 16-bit character strings. It will
600    obtained and extended as required. */
601    
602    #ifdef SUPPORT_PCRE16
603    static int buffer16_size = 0;
604    static pcre_uint16 *buffer16 = NULL;
605    
606    #ifdef SUPPORT_PCRE8
607    
608    /* We need the table of operator lengths that is used for 16-bit compiling, in
609    order to swap bytes in a pattern for saving/reloading testing. Luckily, the
610    data is defined as a macro. However, we must ensure that LINK_SIZE is adjusted
611    appropriately for the 16-bit world. Just as a safety check, make sure that
612    COMPILE_PCRE16 is *not* set. */
613    
614    #ifdef COMPILE_PCRE16
615    #error COMPILE_PCRE16 must not be set when compiling pcretest.c
616    #endif
617    
618    #if LINK_SIZE == 2
619    #undef LINK_SIZE
620    #define LINK_SIZE 1
621    #elif LINK_SIZE == 3 || LINK_SIZE == 4
622    #undef LINK_SIZE
623    #define LINK_SIZE 2
624    #else
625    #error LINK_SIZE must be either 2, 3, or 4
626    #endif
627    
628    #undef IMM2_SIZE
629    #define IMM2_SIZE 1
630    
631    #endif /* SUPPORT_PCRE8 */
632    
633    static const pcre_uint16 OP_lengths16[] = { OP_LENGTHS };
634    #endif  /* SUPPORT_PCRE16 */
635    
636    /* If we have 8-bit support, default use_pcre16 to false; if there is also
637    16-bit support, it can be changed by an option. If there is no 8-bit support,
638    there must be 16-bit support, so default it to 1. */
639    
640    #ifdef SUPPORT_PCRE8
641    static int use_pcre16 = 0;
642    #else
643    static int use_pcre16 = 1;
644    #endif
645    
646    /* Textual explanations for runtime error codes */
647    
648    static const char *errtexts[] = {
649      NULL,  /* 0 is no error */
650      NULL,  /* NOMATCH is handled specially */
651      "NULL argument passed",
652      "bad option value",
653      "magic number missing",
654      "unknown opcode - pattern overwritten?",
655      "no more memory",
656      NULL,  /* never returned by pcre_exec() or pcre_dfa_exec() */
657      "match limit exceeded",
658      "callout error code",
659      NULL,  /* BADUTF8/16 is handled specially */
660      NULL,  /* BADUTF8/16 offset is handled specially */
661      NULL,  /* PARTIAL is handled specially */
662      "not used - internal error",
663      "internal error - pattern overwritten?",
664      "bad count value",
665      "item unsupported for DFA matching",
666      "backreference condition or recursion test not supported for DFA matching",
667      "match limit not supported for DFA matching",
668      "workspace size exceeded in DFA matching",
669      "too much recursion for DFA matching",
670      "recursion limit exceeded",
671      "not used - internal error",
672      "invalid combination of newline options",
673      "bad offset value",
674      NULL,  /* SHORTUTF8/16 is handled specially */
675      "nested recursion at the same subject position",
676      "JIT stack limit reached",
677      "pattern compiled in wrong mode: 8-bit/16-bit error"
678    };
679    
680    
681    /*************************************************
682    *         Alternate character tables             *
683    *************************************************/
684    
685    /* By default, the "tables" pointer when calling PCRE is set to NULL, thereby
686    using the default tables of the library. However, the T option can be used to
687    select alternate sets of tables, for different kinds of testing. Note also that
688    the L (locale) option also adjusts the tables. */
689    
690    /* This is the set of tables distributed as default with PCRE. It recognizes
691    only ASCII characters. */
692    
693    static const pcre_uint8 tables0[] = {
694    
695    /* This table is a lower casing table. */
696    
697        0,  1,  2,  3,  4,  5,  6,  7,
698        8,  9, 10, 11, 12, 13, 14, 15,
699       16, 17, 18, 19, 20, 21, 22, 23,
700       24, 25, 26, 27, 28, 29, 30, 31,
701       32, 33, 34, 35, 36, 37, 38, 39,
702       40, 41, 42, 43, 44, 45, 46, 47,
703       48, 49, 50, 51, 52, 53, 54, 55,
704       56, 57, 58, 59, 60, 61, 62, 63,
705       64, 97, 98, 99,100,101,102,103,
706      104,105,106,107,108,109,110,111,
707      112,113,114,115,116,117,118,119,
708      120,121,122, 91, 92, 93, 94, 95,
709       96, 97, 98, 99,100,101,102,103,
710      104,105,106,107,108,109,110,111,
711      112,113,114,115,116,117,118,119,
712      120,121,122,123,124,125,126,127,
713      128,129,130,131,132,133,134,135,
714      136,137,138,139,140,141,142,143,
715      144,145,146,147,148,149,150,151,
716      152,153,154,155,156,157,158,159,
717      160,161,162,163,164,165,166,167,
718      168,169,170,171,172,173,174,175,
719      176,177,178,179,180,181,182,183,
720      184,185,186,187,188,189,190,191,
721      192,193,194,195,196,197,198,199,
722      200,201,202,203,204,205,206,207,
723      208,209,210,211,212,213,214,215,
724      216,217,218,219,220,221,222,223,
725      224,225,226,227,228,229,230,231,
726      232,233,234,235,236,237,238,239,
727      240,241,242,243,244,245,246,247,
728      248,249,250,251,252,253,254,255,
729    
730    /* This table is a case flipping table. */
731    
732        0,  1,  2,  3,  4,  5,  6,  7,
733        8,  9, 10, 11, 12, 13, 14, 15,
734       16, 17, 18, 19, 20, 21, 22, 23,
735       24, 25, 26, 27, 28, 29, 30, 31,
736       32, 33, 34, 35, 36, 37, 38, 39,
737       40, 41, 42, 43, 44, 45, 46, 47,
738       48, 49, 50, 51, 52, 53, 54, 55,
739       56, 57, 58, 59, 60, 61, 62, 63,
740       64, 97, 98, 99,100,101,102,103,
741      104,105,106,107,108,109,110,111,
742      112,113,114,115,116,117,118,119,
743      120,121,122, 91, 92, 93, 94, 95,
744       96, 65, 66, 67, 68, 69, 70, 71,
745       72, 73, 74, 75, 76, 77, 78, 79,
746       80, 81, 82, 83, 84, 85, 86, 87,
747       88, 89, 90,123,124,125,126,127,
748      128,129,130,131,132,133,134,135,
749      136,137,138,139,140,141,142,143,
750      144,145,146,147,148,149,150,151,
751      152,153,154,155,156,157,158,159,
752      160,161,162,163,164,165,166,167,
753      168,169,170,171,172,173,174,175,
754      176,177,178,179,180,181,182,183,
755      184,185,186,187,188,189,190,191,
756      192,193,194,195,196,197,198,199,
757      200,201,202,203,204,205,206,207,
758      208,209,210,211,212,213,214,215,
759      216,217,218,219,220,221,222,223,
760      224,225,226,227,228,229,230,231,
761      232,233,234,235,236,237,238,239,
762      240,241,242,243,244,245,246,247,
763      248,249,250,251,252,253,254,255,
764    
765    /* This table contains bit maps for various character classes. Each map is 32
766    bytes long and the bits run from the least significant end of each byte. The
767    classes that have their own maps are: space, xdigit, digit, upper, lower, word,
768    graph, print, punct, and cntrl. Other classes are built from combinations. */
769    
770      0x00,0x3e,0x00,0x00,0x01,0x00,0x00,0x00,
771      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
772      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
773      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
774    
775      0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
776      0x7e,0x00,0x00,0x00,0x7e,0x00,0x00,0x00,
777      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
778      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
779    
780      0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
781      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
782      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
783      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
784    
785      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
786      0xfe,0xff,0xff,0x07,0x00,0x00,0x00,0x00,
787      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
788      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
789    
790      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
791      0x00,0x00,0x00,0x00,0xfe,0xff,0xff,0x07,
792      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
793      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
794    
795      0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03,
796      0xfe,0xff,0xff,0x87,0xfe,0xff,0xff,0x07,
797      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
798      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
799    
800      0x00,0x00,0x00,0x00,0xfe,0xff,0xff,0xff,
801      0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f,
802      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
803      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
804    
805      0x00,0x00,0x00,0x00,0xff,0xff,0xff,0xff,
806      0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f,
807      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
808      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
809    
810      0x00,0x00,0x00,0x00,0xfe,0xff,0x00,0xfc,
811      0x01,0x00,0x00,0xf8,0x01,0x00,0x00,0x78,
812      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
813      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
814    
815      0xff,0xff,0xff,0xff,0x00,0x00,0x00,0x00,
816      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x80,
817      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
818      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
819    
820    /* This table identifies various classes of character by individual bits:
821      0x01   white space character
822      0x02   letter
823      0x04   decimal digit
824      0x08   hexadecimal digit
825      0x10   alphanumeric or '_'
826      0x80   regular expression metacharacter or binary zero
827    */
828    
829      0x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*   0-  7 */
830      0x00,0x01,0x01,0x00,0x01,0x01,0x00,0x00, /*   8- 15 */
831      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*  16- 23 */
832      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*  24- 31 */
833      0x01,0x00,0x00,0x00,0x80,0x00,0x00,0x00, /*    - '  */
834      0x80,0x80,0x80,0x80,0x00,0x00,0x80,0x00, /*  ( - /  */
835      0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c, /*  0 - 7  */
836      0x1c,0x1c,0x00,0x00,0x00,0x00,0x00,0x80, /*  8 - ?  */
837      0x00,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /*  @ - G  */
838      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  H - O  */
839      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  P - W  */
840      0x12,0x12,0x12,0x80,0x80,0x00,0x80,0x10, /*  X - _  */
841      0x00,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /*  ` - g  */
842      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  h - o  */
843      0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /*  p - w  */
844      0x12,0x12,0x12,0x80,0x80,0x00,0x00,0x00, /*  x -127 */
845      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 128-135 */
846      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 136-143 */
847      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 144-151 */
848      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 152-159 */
849      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 160-167 */
850      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 168-175 */
851      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 176-183 */
852      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 184-191 */
853      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 192-199 */
854      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 200-207 */
855      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 208-215 */
856      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 216-223 */
857      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 224-231 */
858      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 232-239 */
859      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 240-247 */
860      0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00};/* 248-255 */
861    
862    /* This is a set of tables that came orginally from a Windows user. It seems to
863    be at least an approximation of ISO 8859. In particular, there are characters
864    greater than 128 that are marked as spaces, letters, etc. */
865    
866    static const pcre_uint8 tables1[] = {
867    0,1,2,3,4,5,6,7,
868    8,9,10,11,12,13,14,15,
869    16,17,18,19,20,21,22,23,
870    24,25,26,27,28,29,30,31,
871    32,33,34,35,36,37,38,39,
872    40,41,42,43,44,45,46,47,
873    48,49,50,51,52,53,54,55,
874    56,57,58,59,60,61,62,63,
875    64,97,98,99,100,101,102,103,
876    104,105,106,107,108,109,110,111,
877    112,113,114,115,116,117,118,119,
878    120,121,122,91,92,93,94,95,
879    96,97,98,99,100,101,102,103,
880    104,105,106,107,108,109,110,111,
881    112,113,114,115,116,117,118,119,
882    120,121,122,123,124,125,126,127,
883    128,129,130,131,132,133,134,135,
884    136,137,138,139,140,141,142,143,
885    144,145,146,147,148,149,150,151,
886    152,153,154,155,156,157,158,159,
887    160,161,162,163,164,165,166,167,
888    168,169,170,171,172,173,174,175,
889    176,177,178,179,180,181,182,183,
890    184,185,186,187,188,189,190,191,
891    224,225,226,227,228,229,230,231,
892    232,233,234,235,236,237,238,239,
893    240,241,242,243,244,245,246,215,
894    248,249,250,251,252,253,254,223,
895    224,225,226,227,228,229,230,231,
896    232,233,234,235,236,237,238,239,
897    240,241,242,243,244,245,246,247,
898    248,249,250,251,252,253,254,255,
899    0,1,2,3,4,5,6,7,
900    8,9,10,11,12,13,14,15,
901    16,17,18,19,20,21,22,23,
902    24,25,26,27,28,29,30,31,
903    32,33,34,35,36,37,38,39,
904    40,41,42,43,44,45,46,47,
905    48,49,50,51,52,53,54,55,
906    56,57,58,59,60,61,62,63,
907    64,97,98,99,100,101,102,103,
908    104,105,106,107,108,109,110,111,
909    112,113,114,115,116,117,118,119,
910    120,121,122,91,92,93,94,95,
911    96,65,66,67,68,69,70,71,
912    72,73,74,75,76,77,78,79,
913    80,81,82,83,84,85,86,87,
914    88,89,90,123,124,125,126,127,
915    128,129,130,131,132,133,134,135,
916    136,137,138,139,140,141,142,143,
917    144,145,146,147,148,149,150,151,
918    152,153,154,155,156,157,158,159,
919    160,161,162,163,164,165,166,167,
920    168,169,170,171,172,173,174,175,
921    176,177,178,179,180,181,182,183,
922    184,185,186,187,188,189,190,191,
923    224,225,226,227,228,229,230,231,
924    232,233,234,235,236,237,238,239,
925    240,241,242,243,244,245,246,215,
926    248,249,250,251,252,253,254,223,
927    192,193,194,195,196,197,198,199,
928    200,201,202,203,204,205,206,207,
929    208,209,210,211,212,213,214,247,
930    216,217,218,219,220,221,222,255,
931    0,62,0,0,1,0,0,0,
932    0,0,0,0,0,0,0,0,
933    32,0,0,0,1,0,0,0,
934    0,0,0,0,0,0,0,0,
935    0,0,0,0,0,0,255,3,
936    126,0,0,0,126,0,0,0,
937    0,0,0,0,0,0,0,0,
938    0,0,0,0,0,0,0,0,
939    0,0,0,0,0,0,255,3,
940    0,0,0,0,0,0,0,0,
941    0,0,0,0,0,0,12,2,
942    0,0,0,0,0,0,0,0,
943    0,0,0,0,0,0,0,0,
944    254,255,255,7,0,0,0,0,
945    0,0,0,0,0,0,0,0,
946    255,255,127,127,0,0,0,0,
947    0,0,0,0,0,0,0,0,
948    0,0,0,0,254,255,255,7,
949    0,0,0,0,0,4,32,4,
950    0,0,0,128,255,255,127,255,
951    0,0,0,0,0,0,255,3,
952    254,255,255,135,254,255,255,7,
953    0,0,0,0,0,4,44,6,
954    255,255,127,255,255,255,127,255,
955    0,0,0,0,254,255,255,255,
956    255,255,255,255,255,255,255,127,
957    0,0,0,0,254,255,255,255,
958    255,255,255,255,255,255,255,255,
959    0,2,0,0,255,255,255,255,
960    255,255,255,255,255,255,255,127,
961    0,0,0,0,255,255,255,255,
962    255,255,255,255,255,255,255,255,
963    0,0,0,0,254,255,0,252,
964    1,0,0,248,1,0,0,120,
965    0,0,0,0,254,255,255,255,
966    0,0,128,0,0,0,128,0,
967    255,255,255,255,0,0,0,0,
968    0,0,0,0,0,0,0,128,
969    255,255,255,255,0,0,0,0,
970    0,0,0,0,0,0,0,0,
971    128,0,0,0,0,0,0,0,
972    0,1,1,0,1,1,0,0,
973    0,0,0,0,0,0,0,0,
974    0,0,0,0,0,0,0,0,
975    1,0,0,0,128,0,0,0,
976    128,128,128,128,0,0,128,0,
977    28,28,28,28,28,28,28,28,
978    28,28,0,0,0,0,0,128,
979    0,26,26,26,26,26,26,18,
980    18,18,18,18,18,18,18,18,
981    18,18,18,18,18,18,18,18,
982    18,18,18,128,128,0,128,16,
983    0,26,26,26,26,26,26,18,
984    18,18,18,18,18,18,18,18,
985    18,18,18,18,18,18,18,18,
986    18,18,18,128,128,0,0,0,
987    0,0,0,0,0,1,0,0,
988    0,0,0,0,0,0,0,0,
989    0,0,0,0,0,0,0,0,
990    0,0,0,0,0,0,0,0,
991    1,0,0,0,0,0,0,0,
992    0,0,18,0,0,0,0,0,
993    0,0,20,20,0,18,0,0,
994    0,20,18,0,0,0,0,0,
995    18,18,18,18,18,18,18,18,
996    18,18,18,18,18,18,18,18,
997    18,18,18,18,18,18,18,0,
998    18,18,18,18,18,18,18,18,
999    18,18,18,18,18,18,18,18,
1000    18,18,18,18,18,18,18,18,
1001    18,18,18,18,18,18,18,0,
1002    18,18,18,18,18,18,18,18
1003    };
1004    
1005    
1006    
1007    
1008    #ifndef HAVE_STRERROR
1009    /*************************************************
1010    *     Provide strerror() for non-ANSI libraries  *
1011    *************************************************/
1012    
1013    /* Some old-fashioned systems still around (e.g. SunOS4) don't have strerror()
1014    in their libraries, but can provide the same facility by this simple
1015    alternative function. */
1016    
1017    extern int   sys_nerr;
1018    extern char *sys_errlist[];
1019    
1020    char *
1021    strerror(int n)
1022    {
1023    if (n < 0 || n >= sys_nerr) return "unknown error number";
1024    return sys_errlist[n];
1025    }
1026    #endif /* HAVE_STRERROR */
1027    
1028    
1029    /*************************************************
1030    *         JIT memory callback                    *
1031    *************************************************/
1032    
1033    static pcre_jit_stack* jit_callback(void *arg)
1034    {
1035    return (pcre_jit_stack *)arg;
1036    }
1037    
1038    
1039    #if !defined NOUTF || defined SUPPORT_PCRE16
1040    /*************************************************
1041    *            Convert UTF-8 string to value       *
1042    *************************************************/
1043    
1044    /* This function takes one or more bytes that represents a UTF-8 character,
1045    and returns the value of the character.
1046    
1047    Argument:
1048      utf8bytes   a pointer to the byte vector
1049      vptr        a pointer to an int to receive the value
1050    
1051    Returns:      >  0 => the number of bytes consumed
1052                  -6 to 0 => malformed UTF-8 character at offset = (-return)
1053    */
1054    
1055    static int
1056    utf82ord(pcre_uint8 *utf8bytes, int *vptr)
1057    {
1058    int c = *utf8bytes++;
1059    int d = c;
1060    int i, j, s;
1061    
1062    for (i = -1; i < 6; i++)               /* i is number of additional bytes */
1063      {
1064      if ((d & 0x80) == 0) break;
1065      d <<= 1;
1066      }
1067    
1068    if (i == -1) { *vptr = c; return 1; }  /* ascii character */
1069    if (i == 0 || i == 6) return 0;        /* invalid UTF-8 */
1070    
1071    /* i now has a value in the range 1-5 */
1072    
1073    s = 6*i;
1074    d = (c & utf8_table3[i]) << s;
1075    
1076    for (j = 0; j < i; j++)
1077      {
1078      c = *utf8bytes++;
1079      if ((c & 0xc0) != 0x80) return -(j+1);
1080      s -= 6;
1081      d |= (c & 0x3f) << s;
1082      }
1083    
1084    /* Check that encoding was the correct unique one */
1085    
1086    for (j = 0; j < utf8_table1_size; j++)
1087      if (d <= utf8_table1[j]) break;
1088    if (j != i) return -(i+1);
1089    
1090    /* Valid value */
1091    
1092    *vptr = d;
1093    return i+1;
1094    }
1095    #endif /* NOUTF || SUPPORT_PCRE16 */
1096    
1097    
1098    
1099    #if !defined NOUTF || defined SUPPORT_PCRE16
1100    /*************************************************
1101    *       Convert character value to UTF-8         *
1102    *************************************************/
1103    
1104    /* This function takes an integer value in the range 0 - 0x7fffffff
1105    and encodes it as a UTF-8 character in 0 to 6 bytes.
1106    
1107    Arguments:
1108      cvalue     the character value
1109      utf8bytes  pointer to buffer for result - at least 6 bytes long
1110    
1111    Returns:     number of characters placed in the buffer
1112    */
1113    
1114    static int
1115    ord2utf8(int cvalue, pcre_uint8 *utf8bytes)
1116    {
1117    register int i, j;
1118    for (i = 0; i < utf8_table1_size; i++)
1119      if (cvalue <= utf8_table1[i]) break;
1120    utf8bytes += i;
1121    for (j = i; j > 0; j--)
1122     {
1123     *utf8bytes-- = 0x80 | (cvalue & 0x3f);
1124     cvalue >>= 6;
1125     }
1126    *utf8bytes = utf8_table2[i] | cvalue;
1127    return i + 1;
1128    }
1129    #endif /* NOUTF || SUPPORT_PCRE16 */
1130    
1131    
1132    
1133    #ifdef SUPPORT_PCRE16
1134    /*************************************************
1135    *         Convert a string to 16-bit             *
1136    *************************************************/
1137    
1138    /* In non-UTF mode, the space needed for a 16-bit string is exactly double the
1139    8-bit size. For a UTF-8 string, the size needed for UTF-16 is no more than
1140    double, because up to 0xffff uses no more than 3 bytes in UTF-8 but possibly 4
1141    in UTF-16. Higher values use 4 bytes in UTF-8 and up to 4 bytes in UTF-16. The
1142    result is always left in buffer16.
1143    
1144    Note that this function does not object to surrogate values. This is
1145    deliberate; it makes it possible to construct UTF-16 strings that are invalid,
1146    for the purpose of testing that they are correctly faulted.
1147    
1148    Patterns to be converted are either plain ASCII or UTF-8; data lines are always
1149    in UTF-8 so that values greater than 255 can be handled.
1150    
1151    Arguments:
1152      data       TRUE if converting a data line; FALSE for a regex
1153      p          points to a byte string
1154      utf        true if UTF-8 (to be converted to UTF-16)
1155      len        number of bytes in the string (excluding trailing zero)
1156    
1157    Returns:     number of 16-bit data items used (excluding trailing zero)
1158                 OR -1 if a UTF-8 string is malformed
1159                 OR -2 if a value > 0x10ffff is encountered
1160                 OR -3 if a value > 0xffff is encountered when not in UTF mode
1161    */
1162    
1163    static int
1164    to16(int data, pcre_uint8 *p, int utf, int len)
1165    {
1166    pcre_uint16 *pp;
1167    
1168    if (buffer16_size < 2*len + 2)
1169      {
1170      if (buffer16 != NULL) free(buffer16);
1171      buffer16_size = 2*len + 2;
1172      buffer16 = (pcre_uint16 *)malloc(buffer16_size);
1173      if (buffer16 == NULL)
1174        {
1175        fprintf(stderr, "pcretest: malloc(%d) failed for buffer16\n", buffer16_size);
1176        exit(1);
1177        }
1178      }
1179    
1180    pp = buffer16;
1181    
1182    if (!utf && !data)
1183      {
1184      while (len-- > 0) *pp++ = *p++;
1185      }
1186    
1187    else
1188      {
1189      int c = 0;
1190      while (len > 0)
1191        {
1192        int chlen = utf82ord(p, &c);
1193        if (chlen <= 0) return -1;
1194        if (c > 0x10ffff) return -2;
1195        p += chlen;
1196        len -= chlen;
1197        if (c < 0x10000) *pp++ = c; else
1198          {
1199          if (!utf) return -3;
1200          c -= 0x10000;
1201          *pp++ = 0xD800 | (c >> 10);
1202          *pp++ = 0xDC00 | (c & 0x3ff);
1203          }
1204        }
1205      }
1206    
1207    *pp = 0;
1208    return pp - buffer16;
1209    }
1210    #endif
1211    
1212    
1213  /*************************************************  /*************************************************
# Line 208  Returns:       pointer to the start of n Line 1233  Returns:       pointer to the start of n
1233                 NULL if no data read and EOF reached                 NULL if no data read and EOF reached
1234  */  */
1235    
1236  static uschar *  static pcre_uint8 *
1237  extend_inputline(FILE *f, uschar *start, const char *prompt)  extend_inputline(FILE *f, pcre_uint8 *start, const char *prompt)
1238  {  {
1239  uschar *here = start;  pcre_uint8 *here = start;
1240    
1241  for (;;)  for (;;)
1242    {    {
1243    int rlen = buffer_size - (here - buffer);    int rlen = (int)(buffer_size - (here - buffer));
1244    
1245    if (rlen > 1000)    if (rlen > 1000)
1246      {      {
# Line 245  for (;;) Line 1270  for (;;)
1270      /* Read the next line by normal means, prompting if the file is stdin. */      /* Read the next line by normal means, prompting if the file is stdin. */
1271    
1272        {        {
1273        if (f == stdin) printf(prompt);        if (f == stdin) printf("%s", prompt);
1274        if (fgets((char *)here, rlen,  f) == NULL)        if (fgets((char *)here, rlen,  f) == NULL)
1275          return (here == start)? NULL : start;          return (here == start)? NULL : start;
1276        }        }
# Line 258  for (;;) Line 1283  for (;;)
1283    else    else
1284      {      {
1285      int new_buffer_size = 2*buffer_size;      int new_buffer_size = 2*buffer_size;
1286      uschar *new_buffer = (unsigned char *)malloc(new_buffer_size);      pcre_uint8 *new_buffer = (pcre_uint8 *)malloc(new_buffer_size);
1287      uschar *new_dbuffer = (unsigned char *)malloc(new_buffer_size);      pcre_uint8 *new_dbuffer = (pcre_uint8 *)malloc(new_buffer_size);
1288      uschar *new_pbuffer = (unsigned char *)malloc(new_buffer_size);      pcre_uint8 *new_pbuffer = (pcre_uint8 *)malloc(new_buffer_size);
1289    
1290      if (new_buffer == NULL || new_dbuffer == NULL || new_pbuffer == NULL)      if (new_buffer == NULL || new_dbuffer == NULL || new_pbuffer == NULL)
1291        {        {
# Line 291  return NULL;  /* Control never gets here Line 1316  return NULL;  /* Control never gets here
1316    
1317    
1318    
   
   
   
   
1319  /*************************************************  /*************************************************
1320  *          Read number from string               *  *          Read number from string               *
1321  *************************************************/  *************************************************/
# Line 311  Returns:        the unsigned long Line 1332  Returns:        the unsigned long
1332  */  */
1333    
1334  static int  static int
1335  get_value(unsigned char *str, unsigned char **endptr)  get_value(pcre_uint8 *str, pcre_uint8 **endptr)
1336  {  {
1337  int result = 0;  int result = 0;
1338  while(*str != 0 && isspace(*str)) str++;  while(*str != 0 && isspace(*str)) str++;
# Line 322  return(result); Line 1343  return(result);
1343    
1344    
1345    
   
1346  /*************************************************  /*************************************************
1347  *            Convert UTF-8 string to value       *  *             Print one character                *
1348  *************************************************/  *************************************************/
1349    
1350  /* This function takes one or more bytes that represents a UTF-8 character,  /* Print a single character either literally, or as a hex escape. */
 and returns the value of the character.  
   
 Argument:  
   utf8bytes   a pointer to the byte vector  
   vptr        a pointer to an int to receive the value  
   
 Returns:      >  0 => the number of bytes consumed  
               -6 to 0 => malformed UTF-8 character at offset = (-return)  
 */  
1351    
1352  #if !defined NOUTF8  static int pchar(int c, FILE *f)
   
 static int  
 utf82ord(unsigned char *utf8bytes, int *vptr)  
1353  {  {
1354  int c = *utf8bytes++;  if (PRINTOK(c))
1355  int d = c;    {
1356  int i, j, s;    if (f != NULL) fprintf(f, "%c", c);
1357      return 1;
1358      }
1359    
1360  for (i = -1; i < 6; i++)               /* i is number of additional bytes */  if (c < 0x100)
1361    {    {
1362    if ((d & 0x80) == 0) break;    if (use_utf)
1363    d <<= 1;      {
1364        if (f != NULL) fprintf(f, "\\x{%02x}", c);
1365        return 6;
1366        }
1367      else
1368        {
1369        if (f != NULL) fprintf(f, "\\x%02x", c);
1370        return 4;
1371        }
1372    }    }
1373    
1374  if (i == -1) { *vptr = c; return 1; }  /* ascii character */  if (f != NULL) fprintf(f, "\\x{%02x}", c);
1375  if (i == 0 || i == 6) return 0;        /* invalid UTF-8 */  return (c <= 0x000000ff)? 6 :
1376           (c <= 0x00000fff)? 7 :
1377           (c <= 0x0000ffff)? 8 :
1378           (c <= 0x000fffff)? 9 : 10;
1379    }
1380    
 /* i now has a value in the range 1-5 */  
1381    
 s = 6*i;  
 d = (c & utf8_table3[i]) << s;  
1382    
1383  for (j = 0; j < i; j++)  #ifdef SUPPORT_PCRE8
1384    {  /*************************************************
1385    c = *utf8bytes++;  *         Print 8-bit character string           *
1386    if ((c & 0xc0) != 0x80) return -(j+1);  *************************************************/
   s -= 6;  
   d |= (c & 0x3f) << s;  
   }  
1387    
1388  /* Check that encoding was the correct unique one */  /* Must handle UTF-8 strings in utf8 mode. Yields number of characters printed.
1389    If handed a NULL file, just counts chars without printing. */
1390    
1391  for (j = 0; j < utf8_table1_size; j++)  static int pchars(pcre_uint8 *p, int length, FILE *f)
1392    if (d <= utf8_table1[j]) break;  {
1393  if (j != i) return -(i+1);  int c = 0;
1394    int yield = 0;
1395    
1396  /* Valid value */  if (length < 0)
1397      length = strlen((char *)p);
1398    
1399  *vptr = d;  while (length-- > 0)
1400  return i+1;    {
1401  }  #if !defined NOUTF
1402      if (use_utf)
1403        {
1404        int rc = utf82ord(p, &c);
1405        if (rc > 0 && rc <= length + 1)   /* Mustn't run over the end */
1406          {
1407          length -= rc - 1;
1408          p += rc;
1409          yield += pchar(c, f);
1410          continue;
1411          }
1412        }
1413    #endif
1414      c = *p++;
1415      yield += pchar(c, f);
1416      }
1417    
1418    return yield;
1419    }
1420  #endif  #endif
1421    
1422    
1423    
1424    #ifdef SUPPORT_PCRE16
1425  /*************************************************  /*************************************************
1426  *       Convert character value to UTF-8         *  *    Find length of 0-terminated 16-bit string   *
1427  *************************************************/  *************************************************/
1428    
1429  /* This function takes an integer value in the range 0 - 0x7fffffff  static int strlen16(PCRE_SPTR16 p)
 and encodes it as a UTF-8 character in 0 to 6 bytes.  
   
 Arguments:  
   cvalue     the character value  
   utf8bytes  pointer to buffer for result - at least 6 bytes long  
   
 Returns:     number of characters placed in the buffer  
 */  
   
 #if !defined NOUTF8  
   
 static int  
 ord2utf8(int cvalue, uschar *utf8bytes)  
1430  {  {
1431  register int i, j;  int len = 0;
1432  for (i = 0; i < utf8_table1_size; i++)  while (*p++ != 0) len++;
1433    if (cvalue <= utf8_table1[i]) break;  return len;
 utf8bytes += i;  
 for (j = i; j > 0; j--)  
  {  
  *utf8bytes-- = 0x80 | (cvalue & 0x3f);  
  cvalue >>= 6;  
  }  
 *utf8bytes = utf8_table2[i] | cvalue;  
 return i + 1;  
1434  }  }
1435    #endif  /* SUPPORT_PCRE16 */
 #endif  
   
1436    
1437    
1438    #ifdef SUPPORT_PCRE16
1439  /*************************************************  /*************************************************
1440  *             Print character string             *  *           Print 16-bit character string        *
1441  *************************************************/  *************************************************/
1442    
1443  /* Character string printing function. Must handle UTF-8 strings in utf8  /* Must handle UTF-16 strings in utf mode. Yields number of characters printed.
1444  mode. Yields number of characters printed. If handed a NULL file, just counts  If handed a NULL file, just counts chars without printing. */
 chars without printing. */  
1445    
1446  static int pchars(unsigned char *p, int length, FILE *f)  static int pchars16(PCRE_SPTR16 p, int length, FILE *f)
1447  {  {
 int c = 0;  
1448  int yield = 0;  int yield = 0;
1449    
1450    if (length < 0)
1451      length = strlen16(p);
1452    
1453  while (length-- > 0)  while (length-- > 0)
1454    {    {
1455  #if !defined NOUTF8    int c = *p++ & 0xffff;
1456    if (use_utf8)  #if !defined NOUTF
1457      if (use_utf && c >= 0xD800 && c < 0xDC00 && length > 0)
1458      {      {
1459      int rc = utf82ord(p, &c);      int d = *p & 0xffff;
1460        if (d >= 0xDC00 && d < 0xDFFF)
     if (rc > 0 && rc <= length + 1)   /* Mustn't run over the end */  
1461        {        {
1462        length -= rc - 1;        c = ((c & 0x3ff) << 10) + (d & 0x3ff) + 0x10000;
1463        p += rc;        length--;
1464        if (PRINTHEX(c))        p++;
         {  
         if (f != NULL) fprintf(f, "%c", c);  
         yield++;  
         }  
       else  
         {  
         int n = 4;  
         if (f != NULL) fprintf(f, "\\x{%02x}", c);  
         yield += (n <= 0x000000ff)? 2 :  
                  (n <= 0x00000fff)? 3 :  
                  (n <= 0x0000ffff)? 4 :  
                  (n <= 0x000fffff)? 5 : 6;  
         }  
       continue;  
1465        }        }
1466      }      }
1467  #endif  #endif
1468      yield += pchar(c, f);
1469      }
1470    
1471     /* Not UTF-8, or malformed UTF-8  */  return yield;
1472    }
1473    #endif  /* SUPPORT_PCRE16 */
1474    
1475    c = *p++;  
1476    if (PRINTHEX(c))  
1477      {  #ifdef SUPPORT_PCRE8
1478      if (f != NULL) fprintf(f, "%c", c);  /*************************************************
1479      yield++;  *     Read a capture name (8-bit) and check it   *
1480      }  *************************************************/
1481    else  
1482      {  static pcre_uint8 *
1483      if (f != NULL) fprintf(f, "\\x%02x", c);  read_capture_name8(pcre_uint8 *p, pcre_uint8 **pp, pcre *re)
1484      yield += 4;  {
1485      }  pcre_uint8 *npp = *pp;
1486    while (isalnum(*p)) *npp++ = *p++;
1487    *npp++ = 0;
1488    *npp = 0;
1489    if (pcre_get_stringnumber(re, (char *)(*pp)) < 0)
1490      {
1491      fprintf(outfile, "no parentheses with name \"");
1492      PCHARSV(*pp, 0, -1, outfile);
1493      fprintf(outfile, "\"\n");
1494    }    }
1495    
1496  return yield;  *pp = npp;
1497    return p;
1498    }
1499    #endif  /* SUPPORT_PCRE8 */
1500    
1501    
1502    
1503    #ifdef SUPPORT_PCRE16
1504    /*************************************************
1505    *     Read a capture name (16-bit) and check it  *
1506    *************************************************/
1507    
1508    /* Note that the text being read is 8-bit. */
1509    
1510    static pcre_uint8 *
1511    read_capture_name16(pcre_uint8 *p, pcre_uint16 **pp, pcre *re)
1512    {
1513    pcre_uint16 *npp = *pp;
1514    while (isalnum(*p)) *npp++ = *p++;
1515    *npp++ = 0;
1516    *npp = 0;
1517    if (pcre16_get_stringnumber(re, (PCRE_SPTR16)(*pp)) < 0)
1518      {
1519      fprintf(outfile, "no parentheses with name \"");
1520      PCHARSV(*pp, 0, -1, outfile);
1521      fprintf(outfile, "\"\n");
1522      }
1523    *pp = npp;
1524    return p;
1525  }  }
1526    #endif  /* SUPPORT_PCRE16 */
1527    
1528    
1529    
# Line 509  if (callout_extra) Line 1552  if (callout_extra)
1552      else      else
1553        {        {
1554        fprintf(f, "%2d: ", i/2);        fprintf(f, "%2d: ", i/2);
1555        (void)pchars((unsigned char *)cb->subject + cb->offset_vector[i],        PCHARSV(cb->subject, cb->offset_vector[i],
1556          cb->offset_vector[i+1] - cb->offset_vector[i], f);          cb->offset_vector[i+1] - cb->offset_vector[i], f);
1557        fprintf(f, "\n");        fprintf(f, "\n");
1558        }        }
# Line 522  printed lengths of the substrings. */ Line 1565  printed lengths of the substrings. */
1565    
1566  if (f != NULL) fprintf(f, "--->");  if (f != NULL) fprintf(f, "--->");
1567    
1568  pre_start = pchars((unsigned char *)cb->subject, cb->start_match, f);  PCHARS(pre_start, cb->subject, 0, cb->start_match, f);
1569  post_start = pchars((unsigned char *)(cb->subject + cb->start_match),  PCHARS(post_start, cb->subject, cb->start_match,
1570    cb->current_position - cb->start_match, f);    cb->current_position - cb->start_match, f);
1571    
1572  subject_length = pchars((unsigned char *)cb->subject, cb->subject_length, NULL);  PCHARS(subject_length, cb->subject, 0, cb->subject_length, NULL);
1573    
1574  (void)pchars((unsigned char *)(cb->subject + cb->current_position),  PCHARSV(cb->subject, cb->current_position,
1575    cb->subject_length - cb->current_position, f);    cb->subject_length - cb->current_position, f);
1576    
1577  if (f != NULL) fprintf(f, "\n");  if (f != NULL) fprintf(f, "\n");
# Line 565  fprintf(outfile, "%.*s", (cb->next_item_ Line 1608  fprintf(outfile, "%.*s", (cb->next_item_
1608  fprintf(outfile, "\n");  fprintf(outfile, "\n");
1609  first_callout = 0;  first_callout = 0;
1610    
1611    if (cb->mark != last_callout_mark)
1612      {
1613      if (cb->mark == NULL)
1614        fprintf(outfile, "Latest Mark: <unset>\n");
1615      else
1616        {
1617        fprintf(outfile, "Latest Mark: ");
1618        PCHARSV(cb->mark, 0, -1, outfile);
1619        putc('\n', outfile);
1620        }
1621      last_callout_mark = cb->mark;
1622      }
1623    
1624  if (cb->callout_data != NULL)  if (cb->callout_data != NULL)
1625    {    {
1626    int callout_data = *((int *)(cb->callout_data));    int callout_data = *((int *)(cb->callout_data));
# Line 584  return (cb->callout_number != callout_fa Line 1640  return (cb->callout_number != callout_fa
1640  *            Local malloc functions              *  *            Local malloc functions              *
1641  *************************************************/  *************************************************/
1642    
1643  /* Alternative malloc function, to test functionality and show the size of the  /* Alternative malloc function, to test functionality and save the size of a
1644  compiled re. */  compiled re, which is the first store request that pcre_compile() makes. The
1645    show_malloc variable is set only during matching. */
1646    
1647  static void *new_malloc(size_t size)  static void *new_malloc(size_t size)
1648  {  {
1649  void *block = malloc(size);  void *block = malloc(size);
1650  gotten_store = size;  gotten_store = size;
1651    if (first_gotten_store == 0) first_gotten_store = size;
1652  if (show_malloc)  if (show_malloc)
1653    fprintf(outfile, "malloc       %3d %p\n", (int)size, block);    fprintf(outfile, "malloc       %3d %p\n", (int)size, block);
1654  return block;  return block;
# Line 603  if (show_malloc) Line 1661  if (show_malloc)
1661  free(block);  free(block);
1662  }  }
1663    
   
1664  /* For recursion malloc/free, to test stacking calls */  /* For recursion malloc/free, to test stacking calls */
1665    
1666  static void *stack_malloc(size_t size)  static void *stack_malloc(size_t size)
# Line 626  free(block); Line 1683  free(block);
1683  *          Call pcre_fullinfo()                  *  *          Call pcre_fullinfo()                  *
1684  *************************************************/  *************************************************/
1685    
1686  /* Get one piece of information from the pcre_fullinfo() function */  /* Get one piece of information from the pcre_fullinfo() function. When only
1687    one of 8-bit or 16-bit is supported, use_pcre16 should always have the correct
1688    value, but the code is defensive.
1689    
1690    Arguments:
1691      re        compiled regex
1692      study     study data
1693      option    PCRE_INFO_xxx option
1694      ptr       where to put the data
1695    
1696  static void new_info(pcre *re, pcre_extra *study, int option, void *ptr)  Returns:    0 when OK, < 0 on error
1697    */
1698    
1699    static int
1700    new_info(pcre *re, pcre_extra *study, int option, void *ptr)
1701  {  {
1702  int rc;  int rc;
1703  if ((rc = pcre_fullinfo(re, study, option, ptr)) < 0)  
1704    fprintf(outfile, "Error %d from pcre_fullinfo(%d)\n", rc, option);  if (use_pcre16)
1705    #ifdef SUPPORT_PCRE16
1706      rc = pcre16_fullinfo(re, study, option, ptr);
1707    #else
1708      rc = PCRE_ERROR_BADMODE;
1709    #endif
1710    else
1711    #ifdef SUPPORT_PCRE8
1712      rc = pcre_fullinfo(re, study, option, ptr);
1713    #else
1714      rc = PCRE_ERROR_BADMODE;
1715    #endif
1716    
1717    if (rc < 0)
1718      {
1719      fprintf(outfile, "Error %d from pcre%s_fullinfo(%d)\n", rc,
1720        use_pcre16? "16" : "", option);
1721      if (rc == PCRE_ERROR_BADMODE)
1722        fprintf(outfile, "Running in %s-bit mode but pattern was compiled in "
1723          "%s-bit mode\n", use_pcre16? "16":"8", use_pcre16? "8":"16");
1724      }
1725    
1726    return rc;
1727  }  }
1728    
1729    
1730    
1731  /*************************************************  /*************************************************
1732  *         Byte flipping function                 *  *             Swap byte functions                *
1733  *************************************************/  *************************************************/
1734    
1735  static unsigned long int  /* The following functions swap the bytes of a pcre_uint16 and pcre_uint32
1736  byteflip(unsigned long int value, int n)  value, respectively.
1737    
1738    Arguments:
1739      value        any number
1740    
1741    Returns:       the byte swapped value
1742    */
1743    
1744    static pcre_uint32
1745    swap_uint32(pcre_uint32 value)
1746  {  {
 if (n == 2) return ((value & 0x00ff) << 8) | ((value & 0xff00) >> 8);  
1747  return ((value & 0x000000ff) << 24) |  return ((value & 0x000000ff) << 24) |
1748         ((value & 0x0000ff00) <<  8) |         ((value & 0x0000ff00) <<  8) |
1749         ((value & 0x00ff0000) >>  8) |         ((value & 0x00ff0000) >>  8) |
1750         ((value & 0xff000000) >> 24);         (value >> 24);
1751  }  }
1752    
1753    static pcre_uint16
1754    swap_uint16(pcre_uint16 value)
1755    {
1756    return (value >> 8) | (value << 8);
1757    }
1758    
1759    
1760    
1761    /*************************************************
1762    *        Flip bytes in a compiled pattern        *
1763    *************************************************/
1764    
1765    /* This function is called if the 'F' option was present on a pattern that is
1766    to be written to a file. We flip the bytes of all the integer fields in the
1767    regex data block and the study block. In 16-bit mode this also flips relevant
1768    bytes in the pattern itself. This is to make it possible to test PCRE's
1769    ability to reload byte-flipped patterns, e.g. those compiled on a different
1770    architecture. */
1771    
1772    static void
1773    regexflip(pcre *ere, pcre_extra *extra)
1774    {
1775    real_pcre *re = (real_pcre *)ere;
1776    #ifdef SUPPORT_PCRE16
1777    int op;
1778    pcre_uint16 *ptr = (pcre_uint16 *)re + re->name_table_offset;
1779    int length = re->name_count * re->name_entry_size;
1780    #ifdef SUPPORT_UTF
1781    BOOL utf = (re->options & PCRE_UTF16) != 0;
1782    BOOL utf16_char = FALSE;
1783    #endif /* SUPPORT_UTF */
1784    #endif /* SUPPORT_PCRE16 */
1785    
1786    /* Always flip the bytes in the main data block and study blocks. */
1787    
1788    re->magic_number = REVERSED_MAGIC_NUMBER;
1789    re->size = swap_uint32(re->size);
1790    re->options = swap_uint32(re->options);
1791    re->flags = swap_uint16(re->flags);
1792    re->top_bracket = swap_uint16(re->top_bracket);
1793    re->top_backref = swap_uint16(re->top_backref);
1794    re->first_char = swap_uint16(re->first_char);
1795    re->req_char = swap_uint16(re->req_char);
1796    re->name_table_offset = swap_uint16(re->name_table_offset);
1797    re->name_entry_size = swap_uint16(re->name_entry_size);
1798    re->name_count = swap_uint16(re->name_count);
1799    
1800    if (extra != NULL)
1801      {
1802      pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);
1803      rsd->size = swap_uint32(rsd->size);
1804      rsd->flags = swap_uint32(rsd->flags);
1805      rsd->minlength = swap_uint32(rsd->minlength);
1806      }
1807    
1808    /* In 8-bit mode, that is all we need to do. In 16-bit mode we must swap bytes
1809    in the name table, if present, and then in the pattern itself. */
1810    
1811    #ifdef SUPPORT_PCRE16
1812    if (!use_pcre16) return;
1813    
1814    while(TRUE)
1815      {
1816      /* Swap previous characters. */
1817      while (length-- > 0)
1818        {
1819        *ptr = swap_uint16(*ptr);
1820        ptr++;
1821        }
1822    #ifdef SUPPORT_UTF
1823      if (utf16_char)
1824        {
1825        if ((ptr[-1] & 0xfc00) == 0xd800)
1826          {
1827          /* We know that there is only one extra character in UTF-16. */
1828          *ptr = swap_uint16(*ptr);
1829          ptr++;
1830          }
1831        }
1832      utf16_char = FALSE;
1833    #endif /* SUPPORT_UTF */
1834    
1835      /* Get next opcode. */
1836    
1837      length = 0;
1838      op = *ptr;
1839      *ptr++ = swap_uint16(op);
1840    
1841      switch (op)
1842        {
1843        case OP_END:
1844        return;
1845    
1846    #ifdef SUPPORT_UTF
1847        case OP_CHAR:
1848        case OP_CHARI:
1849        case OP_NOT:
1850        case OP_NOTI:
1851        case OP_STAR:
1852        case OP_MINSTAR:
1853        case OP_PLUS:
1854        case OP_MINPLUS:
1855        case OP_QUERY:
1856        case OP_MINQUERY:
1857        case OP_UPTO:
1858        case OP_MINUPTO:
1859        case OP_EXACT:
1860        case OP_POSSTAR:
1861        case OP_POSPLUS:
1862        case OP_POSQUERY:
1863        case OP_POSUPTO:
1864        case OP_STARI:
1865        case OP_MINSTARI:
1866        case OP_PLUSI:
1867        case OP_MINPLUSI:
1868        case OP_QUERYI:
1869        case OP_MINQUERYI:
1870        case OP_UPTOI:
1871        case OP_MINUPTOI:
1872        case OP_EXACTI:
1873        case OP_POSSTARI:
1874        case OP_POSPLUSI:
1875        case OP_POSQUERYI:
1876        case OP_POSUPTOI:
1877        case OP_NOTSTAR:
1878        case OP_NOTMINSTAR:
1879        case OP_NOTPLUS:
1880        case OP_NOTMINPLUS:
1881        case OP_NOTQUERY:
1882        case OP_NOTMINQUERY:
1883        case OP_NOTUPTO:
1884        case OP_NOTMINUPTO:
1885        case OP_NOTEXACT:
1886        case OP_NOTPOSSTAR:
1887        case OP_NOTPOSPLUS:
1888        case OP_NOTPOSQUERY:
1889        case OP_NOTPOSUPTO:
1890        case OP_NOTSTARI:
1891        case OP_NOTMINSTARI:
1892        case OP_NOTPLUSI:
1893        case OP_NOTMINPLUSI:
1894        case OP_NOTQUERYI:
1895        case OP_NOTMINQUERYI:
1896        case OP_NOTUPTOI:
1897        case OP_NOTMINUPTOI:
1898        case OP_NOTEXACTI:
1899        case OP_NOTPOSSTARI:
1900        case OP_NOTPOSPLUSI:
1901        case OP_NOTPOSQUERYI:
1902        case OP_NOTPOSUPTOI:
1903        if (utf) utf16_char = TRUE;
1904    #endif
1905        /* Fall through. */
1906    
1907        default:
1908        length = OP_lengths16[op] - 1;
1909        break;
1910    
1911        case OP_CLASS:
1912        case OP_NCLASS:
1913        /* Skip the character bit map. */
1914        ptr += 32/sizeof(pcre_uint16);
1915        length = 0;
1916        break;
1917    
1918        case OP_XCLASS:
1919        /* LINK_SIZE can be 1 or 2 in 16 bit mode. */
1920        if (LINK_SIZE > 1)
1921          length = (int)((((unsigned int)(ptr[0]) << 16) | (unsigned int)(ptr[1]))
1922            - (1 + LINK_SIZE + 1));
1923        else
1924          length = (int)((unsigned int)(ptr[0]) - (1 + LINK_SIZE + 1));
1925    
1926        /* Reverse the size of the XCLASS instance. */
1927        *ptr = swap_uint16(*ptr);
1928        ptr++;
1929        if (LINK_SIZE > 1)
1930          {
1931          *ptr = swap_uint16(*ptr);
1932          ptr++;
1933          }
1934    
1935        op = *ptr;
1936        *ptr = swap_uint16(op);
1937        ptr++;
1938        if ((op & XCL_MAP) != 0)
1939          {
1940          /* Skip the character bit map. */
1941          ptr += 32/sizeof(pcre_uint16);
1942          length -= 32/sizeof(pcre_uint16);
1943          }
1944        break;
1945        }
1946      }
1947    /* Control should never reach here in 16 bit mode. */
1948    #endif /* SUPPORT_PCRE16 */
1949    }
1950    
1951    
1952    
# Line 659  return ((value & 0x000000ff) << 24) | Line 1955  return ((value & 0x000000ff) << 24) |
1955  *************************************************/  *************************************************/
1956    
1957  static int  static int
1958  check_match_limit(pcre *re, pcre_extra *extra, uschar *bptr, int len,  check_match_limit(pcre *re, pcre_extra *extra, pcre_uint8 *bptr, int len,
1959    int start_offset, int options, int *use_offsets, int use_size_offsets,    int start_offset, int options, int *use_offsets, int use_size_offsets,
1960    int flag, unsigned long int *limit, int errnumber, const char *msg)    int flag, unsigned long int *limit, int errnumber, const char *msg)
1961  {  {
# Line 674  for (;;) Line 1970  for (;;)
1970    {    {
1971    *limit = mid;    *limit = mid;
1972    
1973    count = pcre_exec(re, extra, (char *)bptr, len, start_offset, options,    PCRE_EXEC(count, re, extra, bptr, len, start_offset, options,
1974      use_offsets, use_size_offsets);      use_offsets, use_size_offsets);
1975    
1976    if (count == errnumber)    if (count == errnumber)
# Line 719  Returns:    < 0, = 0, or > 0, according Line 2015  Returns:    < 0, = 0, or > 0, according
2015  */  */
2016    
2017  static int  static int
2018  strncmpic(uschar *s, uschar *t, int n)  strncmpic(pcre_uint8 *s, pcre_uint8 *t, int n)
2019  {  {
2020  while (n--)  while (n--)
2021    {    {
# Line 735  return 0; Line 2031  return 0;
2031  *         Check newline indicator                *  *         Check newline indicator                *
2032  *************************************************/  *************************************************/
2033    
2034  /* This is used both at compile and run-time to check for <xxx> escapes, where  /* This is used both at compile and run-time to check for <xxx> escapes. Print
2035  xxx is LF, CR, CRLF, ANYCRLF, or ANY. Print a message and return 0 if there is  a message and return 0 if there is no match.
 no match.  
2036    
2037  Arguments:  Arguments:
2038    p           points after the leading '<'    p           points after the leading '<'
# Line 747  Returns:      appropriate PCRE_NEWLINE_x Line 2042  Returns:      appropriate PCRE_NEWLINE_x
2042  */  */
2043    
2044  static int  static int
2045  check_newline(uschar *p, FILE *f)  check_newline(pcre_uint8 *p, FILE *f)
2046  {  {
2047  if (strncmpic(p, (uschar *)"cr>", 3) == 0) return PCRE_NEWLINE_CR;  if (strncmpic(p, (pcre_uint8 *)"cr>", 3) == 0) return PCRE_NEWLINE_CR;
2048  if (strncmpic(p, (uschar *)"lf>", 3) == 0) return PCRE_NEWLINE_LF;  if (strncmpic(p, (pcre_uint8 *)"lf>", 3) == 0) return PCRE_NEWLINE_LF;
2049  if (strncmpic(p, (uschar *)"crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;  if (strncmpic(p, (pcre_uint8 *)"crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;
2050  if (strncmpic(p, (uschar *)"anycrlf>", 8) == 0) return PCRE_NEWLINE_ANYCRLF;  if (strncmpic(p, (pcre_uint8 *)"anycrlf>", 8) == 0) return PCRE_NEWLINE_ANYCRLF;
2051  if (strncmpic(p, (uschar *)"any>", 4) == 0) return PCRE_NEWLINE_ANY;  if (strncmpic(p, (pcre_uint8 *)"any>", 4) == 0) return PCRE_NEWLINE_ANY;
2052  if (strncmpic(p, (uschar *)"bsr_anycrlf>", 12) == 0) return PCRE_BSR_ANYCRLF;  if (strncmpic(p, (pcre_uint8 *)"bsr_anycrlf>", 12) == 0) return PCRE_BSR_ANYCRLF;
2053  if (strncmpic(p, (uschar *)"bsr_unicode>", 12) == 0) return PCRE_BSR_UNICODE;  if (strncmpic(p, (pcre_uint8 *)"bsr_unicode>", 12) == 0) return PCRE_BSR_UNICODE;
2054  fprintf(f, "Unknown newline type at: <%s\n", p);  fprintf(f, "Unknown newline type at: <%s\n", p);
2055  return 0;  return 0;
2056  }  }
# Line 777  printf("If input is a terminal, readline Line 2072  printf("If input is a terminal, readline
2072  printf("This version of pcretest is not linked with readline().\n");  printf("This version of pcretest is not linked with readline().\n");
2073  #endif  #endif
2074  printf("\nOptions:\n");  printf("\nOptions:\n");
2075    #ifdef SUPPORT_PCRE16
2076    printf("  -16      use 16-bit interface\n");
2077    #endif
2078  printf("  -b       show compiled code (bytecode)\n");  printf("  -b       show compiled code (bytecode)\n");
2079  printf("  -C       show PCRE compile-time options and exit\n");  printf("  -C       show PCRE compile-time options and exit\n");
2080    printf("  -C arg   show a specific compile-time option\n");
2081    printf("           and exit with its value. The arg can be:\n");
2082    printf("     linksize     internal link size [2, 3, 4]\n");
2083    printf("     pcre8        8 bit library support enabled [0, 1]\n");
2084    printf("     pcre16       16 bit library support enabled [0, 1]\n");
2085    printf("     utf          Unicode Transformation Format supported [0, 1]\n");
2086    printf("     ucp          Unicode Properties supported [0, 1]\n");
2087    printf("     jit          Just-in-time compiler supported [0, 1]\n");
2088    printf("     newline      Newline type [CR, LF, CRLF, ANYCRLF, ANY, ???]\n");
2089  printf("  -d       debug: show compiled code and information (-b and -i)\n");  printf("  -d       debug: show compiled code and information (-b and -i)\n");
2090  #if !defined NODFA  #if !defined NODFA
2091  printf("  -dfa     force DFA matching for all subjects\n");  printf("  -dfa     force DFA matching for all subjects\n");
2092  #endif  #endif
2093  printf("  -help    show usage information\n");  printf("  -help    show usage information\n");
2094  printf("  -i       show information about compiled patterns\n"  printf("  -i       show information about compiled patterns\n"
2095           "  -M       find MATCH_LIMIT minimum for each subject\n"
2096         "  -m       output memory used information\n"         "  -m       output memory used information\n"
2097         "  -o <n>   set size of offsets vector to <n>\n");         "  -o <n>   set size of offsets vector to <n>\n");
2098  #if !defined NOPOSIX  #if !defined NOPOSIX
# Line 792  printf("  -p       use POSIX interface\n Line 2100  printf("  -p       use POSIX interface\n
2100  #endif  #endif
2101  printf("  -q       quiet: do not output PCRE version number at start\n");  printf("  -q       quiet: do not output PCRE version number at start\n");
2102  printf("  -S <n>   set stack size to <n> megabytes\n");  printf("  -S <n>   set stack size to <n> megabytes\n");
2103  printf("  -s       output store (memory) used information\n"  printf("  -s       force each pattern to be studied at basic level\n"
2104           "  -s+      force each pattern to be studied, using JIT if available\n"
2105         "  -t       time compilation and execution\n");         "  -t       time compilation and execution\n");
2106  printf("  -t <n>   time compilation and execution, repeating <n> times\n");  printf("  -t <n>   time compilation and execution, repeating <n> times\n");
2107  printf("  -tm      time execution (matching) only\n");  printf("  -tm      time execution (matching) only\n");
# Line 812  options, followed by a set of test data, Line 2121  options, followed by a set of test data,
2121  int main(int argc, char **argv)  int main(int argc, char **argv)
2122  {  {
2123  FILE *infile = stdin;  FILE *infile = stdin;
2124    const char *version;
2125  int options = 0;  int options = 0;
2126  int study_options = 0;  int study_options = 0;
2127    int default_find_match_limit = FALSE;
2128  int op = 1;  int op = 1;
2129  int timeit = 0;  int timeit = 0;
2130  int timeitm = 0;  int timeitm = 0;
2131  int showinfo = 0;  int showinfo = 0;
2132  int showstore = 0;  int showstore = 0;
2133    int force_study = -1;
2134    int force_study_options = 0;
2135  int quiet = 0;  int quiet = 0;
2136  int size_offsets = 45;  int size_offsets = 45;
2137  int size_offsets_max;  int size_offsets_max;
# Line 832  int all_use_dfa = 0; Line 2145  int all_use_dfa = 0;
2145  int yield = 0;  int yield = 0;
2146  int stack_size;  int stack_size;
2147    
2148  /* These vectors store, end-to-end, a list of captured substring names. Assume  pcre_jit_stack *jit_stack = NULL;
 that 1024 is plenty long enough for the few names we'll be testing. */  
   
 uschar copynames[1024];  
 uschar getnames[1024];  
   
 uschar *copynamesptr;  
 uschar *getnamesptr;  
2149    
2150  /* Get buffers from malloc() so that Electric Fence will check their misuse  /* These vectors store, end-to-end, a list of zero-terminated captured
2151  when I am debugging. They grow automatically when very long lines are read. */  substring names, each list itself being terminated by an empty name. Assume
2152    that 1024 is plenty long enough for the few names we'll be testing. It is
2153  buffer = (unsigned char *)malloc(buffer_size);  easiest to keep separate 8-bit and 16-bit versions, using the 16-bit version
2154  dbuffer = (unsigned char *)malloc(buffer_size);  for the actual memory, to ensure alignment. By defining these variables always
2155  pbuffer = (unsigned char *)malloc(buffer_size);  (whether or not 8-bit or 16-bit is supported), we avoid too much mess with
2156    #ifdefs in the code. */
2157    
2158    pcre_uint16 copynames[1024];
2159    pcre_uint16 getnames[1024];
2160    
2161    pcre_uint16 *cn16ptr;
2162    pcre_uint16 *gn16ptr;
2163    
2164    pcre_uint8 *copynames8 = (pcre_uint8 *)copynames;
2165    pcre_uint8 *getnames8 = (pcre_uint8 *)getnames;
2166    pcre_uint8 *cn8ptr;
2167    pcre_uint8 *gn8ptr;
2168    
2169    /* Get buffers from malloc() so that valgrind will check their misuse when
2170    debugging. They grow automatically when very long lines are read. The 16-bit
2171    buffer (buffer16) is obtained only if needed. */
2172    
2173    buffer = (pcre_uint8 *)malloc(buffer_size);
2174    dbuffer = (pcre_uint8 *)malloc(buffer_size);
2175    pbuffer = (pcre_uint8 *)malloc(buffer_size);
2176    
2177  /* The outfile variable is static so that new_malloc can use it. */  /* The outfile variable is static so that new_malloc can use it. */
2178    
# Line 861  it set 0x8000, but then I was advised th Line 2187  it set 0x8000, but then I was advised th
2187  _setmode( _fileno( stdout ), _O_BINARY );  _setmode( _fileno( stdout ), _O_BINARY );
2188  #endif  #endif
2189    
2190    /* Get the version number: both pcre_version() and pcre16_version() give the
2191    same answer. We just need to ensure that we call one that is available. */
2192    
2193    #ifdef SUPPORT_PCRE8
2194    version = pcre_version();
2195    #else
2196    version = pcre16_version();
2197    #endif
2198    
2199  /* Scan options */  /* Scan options */
2200    
2201  while (argc > 1 && argv[op][0] == '-')  while (argc > 1 && argv[op][0] == '-')
2202    {    {
2203    unsigned char *endptr;    pcre_uint8 *endptr;
2204    
2205    if (strcmp(argv[op], "-s") == 0 || strcmp(argv[op], "-m") == 0)    if (strcmp(argv[op], "-m") == 0) showstore = 1;
2206      showstore = 1;    else if (strcmp(argv[op], "-s") == 0) force_study = 0;
2207      else if (strcmp(argv[op], "-s+") == 0)
2208        {
2209        force_study = 1;
2210        force_study_options = PCRE_STUDY_JIT_COMPILE;
2211        }
2212      else if (strcmp(argv[op], "-16") == 0)
2213        {
2214    #ifdef SUPPORT_PCRE16
2215        use_pcre16 = 1;
2216    #else
2217        printf("** This version of PCRE was built without 16-bit support\n");
2218        exit(1);
2219    #endif
2220        }
2221    else if (strcmp(argv[op], "-q") == 0) quiet = 1;    else if (strcmp(argv[op], "-q") == 0) quiet = 1;
2222    else if (strcmp(argv[op], "-b") == 0) debug = 1;    else if (strcmp(argv[op], "-b") == 0) debug = 1;
2223    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;
2224    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;
2225      else if (strcmp(argv[op], "-M") == 0) default_find_match_limit = TRUE;
2226  #if !defined NODFA  #if !defined NODFA
2227    else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1;    else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1;
2228  #endif  #endif
2229    else if (strcmp(argv[op], "-o") == 0 && argc > 2 &&    else if (strcmp(argv[op], "-o") == 0 && argc > 2 &&
2230        ((size_offsets = get_value((unsigned char *)argv[op+1], &endptr)),        ((size_offsets = get_value((pcre_uint8 *)argv[op+1], &endptr)),
2231          *endptr == 0))          *endptr == 0))
2232      {      {
2233      op++;      op++;
# Line 887  while (argc > 1 && argv[op][0] == '-') Line 2237  while (argc > 1 && argv[op][0] == '-')
2237      {      {
2238      int both = argv[op][2] == 0;      int both = argv[op][2] == 0;
2239      int temp;      int temp;
2240      if (argc > 2 && (temp = get_value((unsigned char *)argv[op+1], &endptr),      if (argc > 2 && (temp = get_value((pcre_uint8 *)argv[op+1], &endptr),
2241                       *endptr == 0))                       *endptr == 0))
2242        {        {
2243        timeitm = temp;        timeitm = temp;
# Line 898  while (argc > 1 && argv[op][0] == '-') Line 2248  while (argc > 1 && argv[op][0] == '-')
2248      if (both) timeit = timeitm;      if (both) timeit = timeitm;
2249      }      }
2250    else if (strcmp(argv[op], "-S") == 0 && argc > 2 &&    else if (strcmp(argv[op], "-S") == 0 && argc > 2 &&
2251        ((stack_size = get_value((unsigned char *)argv[op+1], &endptr)),        ((stack_size = get_value((pcre_uint8 *)argv[op+1], &endptr)),
2252          *endptr == 0))          *endptr == 0))
2253      {      {
2254  #if defined(_WIN32) || defined(WIN32)  #if defined(_WIN32) || defined(WIN32) || defined(__minix)
2255      printf("PCRE: -S not supported on this OS\n");      printf("PCRE: -S not supported on this OS\n");
2256      exit(1);      exit(1);
2257  #else  #else
# Line 925  while (argc > 1 && argv[op][0] == '-') Line 2275  while (argc > 1 && argv[op][0] == '-')
2275    else if (strcmp(argv[op], "-C") == 0)    else if (strcmp(argv[op], "-C") == 0)
2276      {      {
2277      int rc;      int rc;
2278      printf("PCRE version %s\n", pcre_version());      unsigned long int lrc;
2279    
2280        if (argc > 2)
2281          {
2282          if (strcmp(argv[op + 1], "linksize") == 0)
2283            {
2284            (void)PCRE_CONFIG(PCRE_CONFIG_LINK_SIZE, &rc);
2285            printf("%d\n", rc);
2286            yield = rc;
2287            goto EXIT;
2288            }
2289          if (strcmp(argv[op + 1], "pcre8") == 0)
2290            {
2291    #ifdef SUPPORT_PCRE8
2292            printf("1\n");
2293            yield = 1;
2294    #else
2295            printf("0\n");
2296            yield = 0;
2297    #endif
2298            goto EXIT;
2299            }
2300          if (strcmp(argv[op + 1], "pcre16") == 0)
2301            {
2302    #ifdef SUPPORT_PCRE16
2303            printf("1\n");
2304            yield = 1;
2305    #else
2306            printf("0\n");
2307            yield = 0;
2308    #endif
2309            goto EXIT;
2310            }
2311          if (strcmp(argv[op + 1], "utf") == 0)
2312            {
2313    #ifdef SUPPORT_PCRE8
2314            (void)pcre_config(PCRE_CONFIG_UTF8, &rc);
2315            printf("%d\n", rc);
2316            yield = rc;
2317    #else
2318            (void)pcre16_config(PCRE_CONFIG_UTF16, &rc);
2319            printf("%d\n", rc);
2320            yield = rc;
2321    #endif
2322            goto EXIT;
2323            }
2324          if (strcmp(argv[op + 1], "ucp") == 0)
2325            {
2326            (void)PCRE_CONFIG(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);
2327            printf("%d\n", rc);
2328            yield = rc;
2329            goto EXIT;
2330            }
2331          if (strcmp(argv[op + 1], "jit") == 0)
2332            {
2333            (void)PCRE_CONFIG(PCRE_CONFIG_JIT, &rc);
2334            printf("%d\n", rc);
2335            yield = rc;
2336            goto EXIT;
2337            }
2338          if (strcmp(argv[op + 1], "newline") == 0)
2339            {
2340            (void)PCRE_CONFIG(PCRE_CONFIG_NEWLINE, &rc);
2341            /* Note that these values are always the ASCII values, even
2342            in EBCDIC environments. CR is 13 and NL is 10. */
2343            printf("%s\n", (rc == 13)? "CR" :
2344              (rc == 10)? "LF" : (rc == (13<<8 | 10))? "CRLF" :
2345              (rc == -2)? "ANYCRLF" :
2346              (rc == -1)? "ANY" : "???");
2347            goto EXIT;
2348            }
2349          printf("Unknown -C option: %s\n", argv[op + 1]);
2350          goto EXIT;
2351          }
2352    
2353        printf("PCRE version %s\n", version);
2354      printf("Compiled with\n");      printf("Compiled with\n");
2355    
2356    /* At least one of SUPPORT_PCRE8 and SUPPORT_PCRE16 will be set. If both
2357    are set, either both UTFs are supported or both are not supported. */
2358    
2359    #if defined SUPPORT_PCRE8 && defined SUPPORT_PCRE16
2360        printf("  8-bit and 16-bit support\n");
2361        (void)pcre_config(PCRE_CONFIG_UTF8, &rc);
2362        if (rc)
2363          printf("  UTF-8 and UTF-16 support\n");
2364        else
2365          printf("  No UTF-8 or UTF-16 support\n");
2366    #elif defined SUPPORT_PCRE8
2367        printf("  8-bit support only\n");
2368      (void)pcre_config(PCRE_CONFIG_UTF8, &rc);      (void)pcre_config(PCRE_CONFIG_UTF8, &rc);
2369      printf("  %sUTF-8 support\n", rc? "" : "No ");      printf("  %sUTF-8 support\n", rc? "" : "No ");
2370      (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);  #else
2371        printf("  16-bit support only\n");
2372        (void)pcre16_config(PCRE_CONFIG_UTF16, &rc);
2373        printf("  %sUTF-16 support\n", rc? "" : "No ");
2374    #endif
2375    
2376        (void)PCRE_CONFIG(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);
2377      printf("  %sUnicode properties support\n", rc? "" : "No ");      printf("  %sUnicode properties support\n", rc? "" : "No ");
2378      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);      (void)PCRE_CONFIG(PCRE_CONFIG_JIT, &rc);
2379      printf("  Newline sequence is %s\n", (rc == '\r')? "CR" :      if (rc)
2380        (rc == '\n')? "LF" : (rc == ('\r'<<8 | '\n'))? "CRLF" :        printf("  Just-in-time compiler support\n");
2381        else
2382          printf("  No just-in-time compiler support\n");
2383        (void)PCRE_CONFIG(PCRE_CONFIG_NEWLINE, &rc);
2384        /* Note that these values are always the ASCII values, even
2385        in EBCDIC environments. CR is 13 and NL is 10. */
2386        printf("  Newline sequence is %s\n", (rc == 13)? "CR" :
2387          (rc == 10)? "LF" : (rc == (13<<8 | 10))? "CRLF" :
2388        (rc == -2)? "ANYCRLF" :        (rc == -2)? "ANYCRLF" :
2389        (rc == -1)? "ANY" : "???");        (rc == -1)? "ANY" : "???");
2390      (void)pcre_config(PCRE_CONFIG_BSR, &rc);      (void)PCRE_CONFIG(PCRE_CONFIG_BSR, &rc);
2391      printf("  \\R matches %s\n", rc? "CR, LF, or CRLF only" :      printf("  \\R matches %s\n", rc? "CR, LF, or CRLF only" :
2392                                       "all Unicode newlines");                                       "all Unicode newlines");
2393      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);      (void)PCRE_CONFIG(PCRE_CONFIG_LINK_SIZE, &rc);
2394      printf("  Internal link size = %d\n", rc);      printf("  Internal link size = %d\n", rc);
2395      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);      (void)PCRE_CONFIG(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);
2396      printf("  POSIX malloc threshold = %d\n", rc);      printf("  POSIX malloc threshold = %d\n", rc);
2397      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &rc);      (void)PCRE_CONFIG(PCRE_CONFIG_MATCH_LIMIT, &lrc);
2398      printf("  Default match limit = %d\n", rc);      printf("  Default match limit = %ld\n", lrc);
2399      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &rc);      (void)PCRE_CONFIG(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &lrc);
2400      printf("  Default recursion depth limit = %d\n", rc);      printf("  Default recursion depth limit = %ld\n", lrc);
2401      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);      (void)PCRE_CONFIG(PCRE_CONFIG_STACKRECURSE, &rc);
2402      printf("  Match recursion uses %s\n", rc? "stack" : "heap");      printf("  Match recursion uses %s\n", rc? "stack" : "heap");
2403      goto EXIT;      goto EXIT;
2404      }      }
# Line 1006  if (argc > 2) Line 2457  if (argc > 2)
2457    
2458  /* Set alternative malloc function */  /* Set alternative malloc function */
2459    
2460    #ifdef SUPPORT_PCRE8
2461  pcre_malloc = new_malloc;  pcre_malloc = new_malloc;
2462  pcre_free = new_free;  pcre_free = new_free;
2463  pcre_stack_malloc = stack_malloc;  pcre_stack_malloc = stack_malloc;
2464  pcre_stack_free = stack_free;  pcre_stack_free = stack_free;
2465    #endif
2466    
2467    #ifdef SUPPORT_PCRE16
2468    pcre16_malloc = new_malloc;
2469    pcre16_free = new_free;
2470    pcre16_stack_malloc = stack_malloc;
2471    pcre16_stack_free = stack_free;
2472    #endif
2473    
2474  /* Heading line unless quiet, then prompt for first regex if stdin */  /* Heading line unless quiet, then prompt for first regex if stdin */
2475    
2476  if (!quiet) fprintf(outfile, "PCRE version %s\n\n", pcre_version());  if (!quiet) fprintf(outfile, "PCRE version %s\n\n", version);
2477    
2478  /* Main loop */  /* Main loop */
2479    
# Line 1028  while (!done) Line 2488  while (!done)
2488  #endif  #endif
2489    
2490    const char *error;    const char *error;
2491    unsigned char *p, *pp, *ppp;    pcre_uint8 *markptr;
2492    unsigned char *to_file = NULL;    pcre_uint8 *p, *pp, *ppp;
2493    const unsigned char *tables = NULL;    pcre_uint8 *to_file = NULL;
2494      const pcre_uint8 *tables = NULL;
2495    unsigned long int true_size, true_study_size = 0;    unsigned long int true_size, true_study_size = 0;
2496    size_t size, regex_gotten_store;    size_t size, regex_gotten_store;
2497      int do_allcaps = 0;
2498      int do_mark = 0;
2499    int do_study = 0;    int do_study = 0;
2500      int no_force_study = 0;
2501    int do_debug = debug;    int do_debug = debug;
2502    int do_G = 0;    int do_G = 0;
2503    int do_g = 0;    int do_g = 0;
2504    int do_showinfo = showinfo;    int do_showinfo = showinfo;
2505    int do_showrest = 0;    int do_showrest = 0;
2506      int do_showcaprest = 0;
2507    int do_flip = 0;    int do_flip = 0;
2508    int erroroffset, len, delimiter, poffset;    int erroroffset, len, delimiter, poffset;
2509    
2510    use_utf8 = 0;    use_utf = 0;
2511    debug_lengths = 1;    debug_lengths = 1;
2512    
2513    if (extend_inputline(infile, buffer, "  re> ") == NULL) break;    if (extend_inputline(infile, buffer, "  re> ") == NULL) break;
# Line 1058  while (!done) Line 2523  while (!done)
2523    if (*p == '<' && strchr((char *)(p+1), '<') == NULL)    if (*p == '<' && strchr((char *)(p+1), '<') == NULL)
2524      {      {
2525      unsigned long int magic, get_options;      unsigned long int magic, get_options;
2526      uschar sbuf[8];      pcre_uint8 sbuf[8];
2527      FILE *f;      FILE *f;
2528    
2529      p++;      p++;
2530        if (*p == '!')
2531          {
2532          do_debug = TRUE;
2533          do_showinfo = TRUE;
2534          p++;
2535          }
2536    
2537      pp = p + (int)strlen((char *)p);      pp = p + (int)strlen((char *)p);
2538      while (isspace(pp[-1])) pp--;      while (isspace(pp[-1])) pp--;
2539      *pp = 0;      *pp = 0;
# Line 1073  while (!done) Line 2545  while (!done)
2545        continue;        continue;
2546        }        }
2547    
2548        first_gotten_store = 0;
2549      if (fread(sbuf, 1, 8, f) != 8) goto FAIL_READ;      if (fread(sbuf, 1, 8, f) != 8) goto FAIL_READ;
2550    
2551      true_size =      true_size =
# Line 1081  while (!done) Line 2554  while (!done)
2554        (sbuf[4] << 24) | (sbuf[5] << 16) | (sbuf[6] << 8) | sbuf[7];        (sbuf[4] << 24) | (sbuf[5] << 16) | (sbuf[6] << 8) | sbuf[7];
2555    
2556      re = (real_pcre *)new_malloc(true_size);      re = (real_pcre *)new_malloc(true_size);
2557      regex_gotten_store = gotten_store;      regex_gotten_store = first_gotten_store;
2558    
2559      if (fread(re, 1, true_size, f) != true_size) goto FAIL_READ;      if (fread(re, 1, true_size, f) != true_size) goto FAIL_READ;
2560    
2561      magic = ((real_pcre *)re)->magic_number;      magic = ((real_pcre *)re)->magic_number;
2562      if (magic != MAGIC_NUMBER)      if (magic != MAGIC_NUMBER)
2563        {        {
2564        if (byteflip(magic, sizeof(magic)) == MAGIC_NUMBER)        if (swap_uint32(magic) == MAGIC_NUMBER)
2565          {          {
2566          do_flip = 1;          do_flip = 1;
2567          }          }
# Line 1100  while (!done) Line 2573  while (!done)
2573          }          }
2574        }        }
2575    
2576      fprintf(outfile, "Compiled regex%s loaded from %s\n",      /* We hide the byte-invert info for little and big endian tests. */
2577        do_flip? " (byte-inverted)" : "", p);      fprintf(outfile, "Compiled pattern%s loaded from %s\n",
2578          do_flip && (p[-1] == '<') ? " (byte-inverted)" : "", p);
     /* Need to know if UTF-8 for printing data strings */  
   
     new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);  
     use_utf8 = (get_options & PCRE_UTF8) != 0;  
2579    
2580      /* Now see if there is any following study data */      /* Now see if there is any following study data. */
2581    
2582      if (true_study_size != 0)      if (true_study_size != 0)
2583        {        {
# Line 1122  while (!done) Line 2591  while (!done)
2591    
2592        if (fread(psd, 1, true_study_size, f) != true_study_size)        if (fread(psd, 1, true_study_size, f) != true_study_size)
2593          {          {
2594          FAIL_READ:          FAIL_READ:
2595          fprintf(outfile, "Failed to read data from %s\n", p);          fprintf(outfile, "Failed to read data from %s\n", p);
2596          if (extra != NULL) new_free(extra);          if (extra != NULL)
2597          if (re != NULL) new_free(re);            {
2598          fclose(f);            PCRE_FREE_STUDY(extra);
2599              }
2600            if (re != NULL) new_free(re);
2601            fclose(f);
2602            continue;
2603            }
2604          fprintf(outfile, "Study data loaded from %s\n", p);
2605          do_study = 1;     /* To get the data output if requested */
2606          }
2607        else fprintf(outfile, "No study data\n");
2608    
2609        /* Flip the necessary bytes. */
2610        if (do_flip)
2611          {
2612          int rc;
2613          PCRE_PATTERN_TO_HOST_BYTE_ORDER(rc, re, extra, NULL);
2614          if (rc == PCRE_ERROR_BADMODE)
2615            {
2616            /* Simulate the result of the function call below. */
2617            fprintf(outfile, "Error %d from pcre%s_fullinfo(%d)\n", rc,
2618              use_pcre16? "16" : "", PCRE_INFO_OPTIONS);
2619            fprintf(outfile, "Running in %s-bit mode but pattern was compiled in "
2620              "%s-bit mode\n", use_pcre16? "16":"8", use_pcre16? "8":"16");
2621          continue;          continue;
2622          }          }
       fprintf(outfile, "Study data loaded from %s\n", p);  
       do_study = 1;     /* To get the data output if requested */  
2623        }        }
2624      else fprintf(outfile, "No study data\n");  
2625        /* Need to know if UTF-8 for printing data strings. */
2626    
2627        if (new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options) < 0) continue;
2628        use_utf = (get_options & PCRE_UTF8) != 0;
2629    
2630      fclose(f);      fclose(f);
2631      goto SHOW_INFO;      goto SHOW_INFO;
2632      }      }
2633    
2634    /* In-line pattern (the usual case). Get the delimiter and seek the end of    /* In-line pattern (the usual case). Get the delimiter and seek the end of
2635    the pattern; if is isn't complete, read more. */    the pattern; if it isn't complete, read more. */
2636    
2637    delimiter = *p++;    delimiter = *p++;
2638    
# Line 1150  while (!done) Line 2643  while (!done)
2643      }      }
2644    
2645    pp = p;    pp = p;
2646    poffset = p - buffer;    poffset = (int)(p - buffer);
2647    
2648    for(;;)    for(;;)
2649      {      {
# Line 1204  while (!done) Line 2697  while (!done)
2697        case 's': options |= PCRE_DOTALL; break;        case 's': options |= PCRE_DOTALL; break;
2698        case 'x': options |= PCRE_EXTENDED; break;        case 'x': options |= PCRE_EXTENDED; break;
2699    
2700        case '+': do_showrest = 1; break;        case '+':
2701          if (do_showrest) do_showcaprest = 1; else do_showrest = 1;
2702          break;
2703    
2704          case '=': do_allcaps = 1; break;
2705        case 'A': options |= PCRE_ANCHORED; break;        case 'A': options |= PCRE_ANCHORED; break;
2706        case 'B': do_debug = 1; break;        case 'B': do_debug = 1; break;
2707        case 'C': options |= PCRE_AUTO_CALLOUT; break;        case 'C': options |= PCRE_AUTO_CALLOUT; break;
# Line 1214  while (!done) Line 2711  while (!done)
2711        case 'G': do_G = 1; break;        case 'G': do_G = 1; break;
2712        case 'I': do_showinfo = 1; break;        case 'I': do_showinfo = 1; break;
2713        case 'J': options |= PCRE_DUPNAMES; break;        case 'J': options |= PCRE_DUPNAMES; break;
2714          case 'K': do_mark = 1; break;
2715        case 'M': log_store = 1; break;        case 'M': log_store = 1; break;
2716        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;        case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;
2717    
# Line 1221  while (!done) Line 2719  while (!done)
2719        case 'P': do_posix = 1; break;        case 'P': do_posix = 1; break;
2720  #endif  #endif
2721    
2722        case 'S': do_study = 1; break;        case 'S':
2723          if (do_study == 0)
2724            {
2725            do_study = 1;
2726            if (*pp == '+')
2727              {
2728              study_options |= PCRE_STUDY_JIT_COMPILE;
2729              pp++;
2730              }
2731            }
2732          else
2733            {
2734            do_study = 0;
2735            no_force_study = 1;
2736            }
2737          break;
2738    
2739        case 'U': options |= PCRE_UNGREEDY; break;        case 'U': options |= PCRE_UNGREEDY; break;
2740          case 'W': options |= PCRE_UCP; break;
2741        case 'X': options |= PCRE_EXTRA; break;        case 'X': options |= PCRE_EXTRA; break;
2742          case 'Y': options |= PCRE_NO_START_OPTIMISE; break;
2743        case 'Z': debug_lengths = 0; break;        case 'Z': debug_lengths = 0; break;
2744        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;        case '8': options |= PCRE_UTF8; use_utf = 1; break;
2745        case '?': options |= PCRE_NO_UTF8_CHECK; break;        case '?': options |= PCRE_NO_UTF8_CHECK; break;
2746    
2747          case 'T':
2748          switch (*pp++)
2749            {
2750            case '0': tables = tables0; break;
2751            case '1': tables = tables1; break;
2752    
2753            case '\r':
2754            case '\n':
2755            case ' ':
2756            case 0:
2757            fprintf(outfile, "** Missing table number after /T\n");
2758            goto SKIP_DATA;
2759    
2760            default:
2761            fprintf(outfile, "** Bad table number \"%c\" after /T\n", pp[-1]);
2762            goto SKIP_DATA;
2763            }
2764          break;
2765    
2766        case 'L':        case 'L':
2767        ppp = pp;        ppp = pp;
2768        /* The '\r' test here is so that it works on Windows. */        /* The '\r' test here is so that it works on Windows. */
# Line 1240  while (!done) Line 2775  while (!done)
2775          goto SKIP_DATA;          goto SKIP_DATA;
2776          }          }
2777        locale_set = 1;        locale_set = 1;
2778        tables = pcre_maketables();        tables = PCRE_MAKETABLES;
2779        pp = ppp;        pp = ppp;
2780        break;        break;
2781    
# Line 1253  while (!done) Line 2788  while (!done)
2788    
2789        case '<':        case '<':
2790          {          {
2791          if (strncmp((char *)pp, "JS>", 3) == 0)          if (strncmpic(pp, (pcre_uint8 *)"JS>", 3) == 0)
2792            {            {
2793            options |= PCRE_JAVASCRIPT_COMPAT;            options |= PCRE_JAVASCRIPT_COMPAT;
2794            pp += 3;            pp += 3;
# Line 1281  while (!done) Line 2816  while (!done)
2816    
2817    /* Handle compiling via the POSIX interface, which doesn't support the    /* Handle compiling via the POSIX interface, which doesn't support the
2818    timing, showing, or debugging options, nor the ability to pass over    timing, showing, or debugging options, nor the ability to pass over
2819    local character tables. */    local character tables. Neither does it have 16-bit support. */
2820    
2821  #if !defined NOPOSIX  #if !defined NOPOSIX
2822    if (posix || do_posix)    if (posix || do_posix)
# Line 1294  while (!done) Line 2829  while (!done)
2829      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;
2830      if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;      if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;
2831      if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;      if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;
2832        if ((options & PCRE_UCP) != 0) cflags |= REG_UCP;
2833        if ((options & PCRE_UNGREEDY) != 0) cflags |= REG_UNGREEDY;
2834    
2835        first_gotten_store = 0;
2836      rc = regcomp(&preg, (char *)p, cflags);      rc = regcomp(&preg, (char *)p, cflags);
2837    
2838      /* Compilation failed; go back for another re, skipping to blank line      /* Compilation failed; go back for another re, skipping to blank line
# Line 1314  while (!done) Line 2852  while (!done)
2852  #endif  /* !defined NOPOSIX */  #endif  /* !defined NOPOSIX */
2853    
2854      {      {
2855        unsigned long int get_options;
2856    
2857        /* In 16-bit mode, convert the input. */
2858    
2859    #ifdef SUPPORT_PCRE16
2860        if (use_pcre16)
2861          {
2862          switch(to16(FALSE, p, options & PCRE_UTF8, (int)strlen((char *)p)))
2863            {
2864            case -1:
2865            fprintf(outfile, "**Failed: invalid UTF-8 string cannot be "
2866              "converted to UTF-16\n");
2867            goto SKIP_DATA;
2868    
2869            case -2:
2870            fprintf(outfile, "**Failed: character value greater than 0x10ffff "
2871              "cannot be converted to UTF-16\n");
2872            goto SKIP_DATA;
2873    
2874            case -3: /* "Impossible error" when to16 is called arg1 FALSE */
2875            fprintf(outfile, "**Failed: character value greater than 0xffff "
2876              "cannot be converted to 16-bit in non-UTF mode\n");
2877            goto SKIP_DATA;
2878    
2879            default:
2880            break;
2881            }
2882          p = (pcre_uint8 *)buffer16;
2883          }
2884    #endif
2885    
2886        /* Compile many times when timing */
2887    
2888      if (timeit > 0)      if (timeit > 0)
2889        {        {
2890        register int i;        register int i;
# Line 1321  while (!done) Line 2892  while (!done)
2892        clock_t start_time = clock();        clock_t start_time = clock();
2893        for (i = 0; i < timeit; i++)        for (i = 0; i < timeit; i++)
2894          {          {
2895          re = pcre_compile((char *)p, options, &error, &erroroffset, tables);          PCRE_COMPILE(re, p, options, &error, &erroroffset, tables);
2896          if (re != NULL) free(re);          if (re != NULL) free(re);
2897          }          }
2898        time_taken = clock() - start_time;        time_taken = clock() - start_time;
# Line 1330  while (!done) Line 2901  while (!done)
2901            (double)CLOCKS_PER_SEC);            (double)CLOCKS_PER_SEC);
2902        }        }
2903    
2904      re = pcre_compile((char *)p, options, &error, &erroroffset, tables);      first_gotten_store = 0;
2905        PCRE_COMPILE(re, p, options, &error, &erroroffset, tables);
2906    
2907      /* Compilation failed; go back for another re, skipping to blank line      /* Compilation failed; go back for another re, skipping to blank line
2908      if non-interactive. */      if non-interactive. */
# Line 1357  while (!done) Line 2929  while (!done)
2929        goto CONTINUE;        goto CONTINUE;
2930        }        }
2931    
2932      /* Compilation succeeded; print data if required. There are now two      /* Compilation succeeded. It is now possible to set the UTF-8 option from
2933      info-returning functions. The old one has a limited interface and      within the regex; check for this so that we know how to process the data
2934      returns only limited data. Check that it agrees with the newer one. */      lines. */
2935    
2936      if (log_store)      if (new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options) < 0)
2937        fprintf(outfile, "Memory allocation (code space): %d\n",        goto SKIP_DATA;
2938          (int)(gotten_store -      if ((get_options & PCRE_UTF8) != 0) use_utf = 1;
               sizeof(real_pcre) -  
               ((real_pcre *)re)->name_count * ((real_pcre *)re)->name_entry_size));  
2939    
2940      /* Extract the size for possible writing before possibly flipping it,      /* Extract the size for possible writing before possibly flipping it,
2941      and remember the store that was got. */      and remember the store that was got. */
2942    
2943      true_size = ((real_pcre *)re)->size;      true_size = ((real_pcre *)re)->size;
2944      regex_gotten_store = gotten_store;      regex_gotten_store = first_gotten_store;
2945    
2946        /* Output code size information if requested */
2947    
2948        if (log_store)
2949          fprintf(outfile, "Memory allocation (code space): %d\n",
2950            (int)(first_gotten_store -
2951                  sizeof(real_pcre) -
2952                  ((real_pcre *)re)->name_count * ((real_pcre *)re)->name_entry_size));
2953    
2954      /* If /S was present, study the regexp to generate additional info to      /* If -s or /S was present, study the regex to generate additional info to
2955      help with the matching. */      help with the matching, unless the pattern has the SS option, which
2956        suppresses the effect of /S (used for a few test patterns where studying is
2957        never sensible). */
2958    
2959      if (do_study)      if (do_study || (force_study >= 0 && !no_force_study))
2960        {        {
2961        if (timeit > 0)        if (timeit > 0)
2962          {          {
# Line 1384  while (!done) Line 2964  while (!done)
2964          clock_t time_taken;          clock_t time_taken;
2965          clock_t start_time = clock();          clock_t start_time = clock();
2966          for (i = 0; i < timeit; i++)          for (i = 0; i < timeit; i++)
2967            extra = pcre_study(re, study_options, &error);            {
2968              PCRE_STUDY(extra, re, study_options | force_study_options, &error);
2969              }
2970          time_taken = clock() - start_time;          time_taken = clock() - start_time;
2971          if (extra != NULL) free(extra);          if (extra != NULL)
2972              {
2973              PCRE_FREE_STUDY(extra);
2974              }
2975          fprintf(outfile, "  Study time %.4f milliseconds\n",          fprintf(outfile, "  Study time %.4f milliseconds\n",
2976            (((double)time_taken * 1000.0) / (double)timeit) /            (((double)time_taken * 1000.0) / (double)timeit) /
2977              (double)CLOCKS_PER_SEC);              (double)CLOCKS_PER_SEC);
2978          }          }
2979        extra = pcre_study(re, study_options, &error);        PCRE_STUDY(extra, re, study_options | force_study_options, &error);
2980        if (error != NULL)        if (error != NULL)
2981          fprintf(outfile, "Failed to study: %s\n", error);          fprintf(outfile, "Failed to study: %s\n", error);
2982        else if (extra != NULL)        else if (extra != NULL)
2983            {
2984          true_study_size = ((pcre_study_data *)(extra->study_data))->size;          true_study_size = ((pcre_study_data *)(extra->study_data))->size;
2985            if (log_store)
2986              {
2987              size_t jitsize;
2988              if (new_info(re, extra, PCRE_INFO_JITSIZE, &jitsize) == 0 &&
2989                  jitsize != 0)
2990                fprintf(outfile, "Memory allocation (JIT code): %d\n", (int)jitsize);
2991              }
2992            }
2993        }        }
2994    
2995      /* If the 'F' option was present, we flip the bytes of all the integer      /* If /K was present, we set up for handling MARK data. */
     fields in the regex data block and the study block. This is to make it  
     possible to test PCRE's handling of byte-flipped patterns, e.g. those  
     compiled on a different architecture. */  
2996    
2997      if (do_flip)      if (do_mark)
2998        {        {
2999        real_pcre *rre = (real_pcre *)re;        if (extra == NULL)
       rre->magic_number =  
         byteflip(rre->magic_number, sizeof(rre->magic_number));  
       rre->size = byteflip(rre->size, sizeof(rre->size));  
       rre->options = byteflip(rre->options, sizeof(rre->options));  
       rre->flags = (pcre_uint16)byteflip(rre->flags, sizeof(rre->flags));  
       rre->top_bracket =  
         (pcre_uint16)byteflip(rre->top_bracket, sizeof(rre->top_bracket));  
       rre->top_backref =  
         (pcre_uint16)byteflip(rre->top_backref, sizeof(rre->top_backref));  
       rre->first_byte =  
         (pcre_uint16)byteflip(rre->first_byte, sizeof(rre->first_byte));  
       rre->req_byte =  
         (pcre_uint16)byteflip(rre->req_byte, sizeof(rre->req_byte));  
       rre->name_table_offset = (pcre_uint16)byteflip(rre->name_table_offset,  
         sizeof(rre->name_table_offset));  
       rre->name_entry_size = (pcre_uint16)byteflip(rre->name_entry_size,  
         sizeof(rre->name_entry_size));  
       rre->name_count = (pcre_uint16)byteflip(rre->name_count,  
         sizeof(rre->name_count));  
   
       if (extra != NULL)  
3000          {          {
3001          pcre_study_data *rsd = (pcre_study_data *)(extra->study_data);          extra = (pcre_extra *)malloc(sizeof(pcre_extra));
3002          rsd->size = byteflip(rsd->size, sizeof(rsd->size));          extra->flags = 0;
         rsd->options = byteflip(rsd->options, sizeof(rsd->options));  
3003          }          }
3004          extra->mark = &markptr;
3005          extra->flags |= PCRE_EXTRA_MARK;
3006        }        }
3007    
3008      /* Extract information from the compiled data if required */      /* Extract and display information from the compiled data if required. */
3009    
3010      SHOW_INFO:      SHOW_INFO:
3011    
3012      if (do_debug)      if (do_debug)
3013        {        {
3014        fprintf(outfile, "------------------------------------------------------------------\n");        fprintf(outfile, "------------------------------------------------------------------\n");
3015        pcre_printint(re, outfile, debug_lengths);        PCRE_PRINTINT(re, outfile, debug_lengths);
3016        }        }
3017    
3018        /* We already have the options in get_options (see above) */
3019    
3020      if (do_showinfo)      if (do_showinfo)
3021        {        {
3022        unsigned long int get_options, all_options;        unsigned long int all_options;
 #if !defined NOINFOCHECK  
       int old_first_char, old_options, old_count;  
 #endif  
3023        int count, backrefmax, first_char, need_char, okpartial, jchanged,        int count, backrefmax, first_char, need_char, okpartial, jchanged,
3024          hascrorlf;          hascrorlf;
3025        int nameentrysize, namecount;        int nameentrysize, namecount;
3026        const uschar *nametable;        const pcre_uint8 *nametable;
3027    
3028        new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);        if (new_info(re, NULL, PCRE_INFO_SIZE, &size) +
3029        new_info(re, NULL, PCRE_INFO_SIZE, &size);            new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count) +
3030        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);            new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax) +
3031        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);            new_info(re, NULL, PCRE_INFO_FIRSTBYTE, &first_char) +
3032        new_info(re, NULL, PCRE_INFO_FIRSTBYTE, &first_char);            new_info(re, NULL, PCRE_INFO_LASTLITERAL, &need_char) +
3033        new_info(re, NULL, PCRE_INFO_LASTLITERAL, &need_char);            new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize) +
3034        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);            new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount) +
3035        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);            new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable) +
3036        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);            new_info(re, NULL, PCRE_INFO_OKPARTIAL, &okpartial) +
3037        new_info(re, NULL, PCRE_INFO_OKPARTIAL, &okpartial);            new_info(re, NULL, PCRE_INFO_JCHANGED, &jchanged) +
3038        new_info(re, NULL, PCRE_INFO_JCHANGED, &jchanged);            new_info(re, NULL, PCRE_INFO_HASCRORLF, &hascrorlf)
3039        new_info(re, NULL, PCRE_INFO_HASCRORLF, &hascrorlf);            != 0)
3040            goto SKIP_DATA;
 #if !defined NOINFOCHECK  
       old_count = pcre_info(re, &old_options, &old_first_char);  
       if (count < 0) fprintf(outfile,  
         "Error %d from pcre_info()\n", count);  
       else  
         {  
         if (old_count != count) fprintf(outfile,  
           "Count disagreement: pcre_fullinfo=%d pcre_info=%d\n", count,  
             old_count);  
   
         if (old_first_char != first_char) fprintf(outfile,  
           "First char disagreement: pcre_fullinfo=%d pcre_info=%d\n",  
             first_char, old_first_char);  
   
         if (old_options != (int)get_options) fprintf(outfile,  
           "Options disagreement: pcre_fullinfo=%ld pcre_info=%d\n",  
             get_options, old_options);  
         }  
 #endif  
3041    
3042        if (size != regex_gotten_store) fprintf(outfile,        if (size != regex_gotten_store) fprintf(outfile,
3043          "Size disagreement: pcre_fullinfo=%d call to malloc for %d\n",          "Size disagreement: pcre_fullinfo=%d call to malloc for %d\n",
# Line 1501  while (!done) Line 3052  while (!done)
3052          fprintf(outfile, "Named capturing subpatterns:\n");          fprintf(outfile, "Named capturing subpatterns:\n");
3053          while (namecount-- > 0)          while (namecount-- > 0)
3054            {            {
3055            fprintf(outfile, "  %s %*s%3d\n", nametable + 2,  #if defined SUPPORT_PCRE8 && defined SUPPORT_PCRE16
3056              nameentrysize - 3 - (int)strlen((char *)nametable + 2), "",            int imm2_size = use_pcre16 ? 1 : 2;
3057              GET2(nametable, 0));  #else
3058              int imm2_size = IMM2_SIZE;
3059    #endif
3060              int length = (int)STRLEN(nametable + imm2_size);
3061              fprintf(outfile, "  ");
3062              PCHARSV(nametable, imm2_size, length, outfile);
3063              while (length++ < nameentrysize - imm2_size) putc(' ', outfile);
3064    #if defined SUPPORT_PCRE8 && defined SUPPORT_PCRE16
3065              fprintf(outfile, "%3d\n", use_pcre16?
3066                 (int)(((PCRE_SPTR16)nametable)[0])
3067                :((int)nametable[0] << 8) | (int)nametable[1]);
3068              nametable += nameentrysize * (use_pcre16 ? 2 : 1);
3069    #else
3070              fprintf(outfile, "%3d\n", GET2(nametable, 0));
3071    #ifdef SUPPORT_PCRE8
3072            nametable += nameentrysize;            nametable += nameentrysize;
3073    #else
3074              nametable += nameentrysize * 2;
3075    #endif
3076    #endif
3077            }            }
3078          }          }
3079    
# Line 1512  while (!done) Line 3081  while (!done)
3081        if (hascrorlf) fprintf(outfile, "Contains explicit CR or LF match\n");        if (hascrorlf) fprintf(outfile, "Contains explicit CR or LF match\n");
3082    
3083        all_options = ((real_pcre *)re)->options;        all_options = ((real_pcre *)re)->options;
3084        if (do_flip) all_options = byteflip(all_options, sizeof(all_options));        if (do_flip) all_options = swap_uint32(all_options);
3085    
3086        if (get_options == 0) fprintf(outfile, "No options\n");        if (get_options == 0) fprintf(outfile, "No options\n");
3087          else fprintf(outfile, "Options:%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s\n",          else fprintf(outfile, "Options:%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s\n",
3088            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",
3089            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",
3090            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",
# Line 1528  while (!done) Line 3097  while (!done)
3097            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",
3098            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",
3099            ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",            ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",
3100            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",            ((get_options & PCRE_UTF8) != 0)? " utf" : "",
3101            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",            ((get_options & PCRE_UCP) != 0)? " ucp" : "",
3102              ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf_check" : "",
3103              ((get_options & PCRE_NO_START_OPTIMIZE) != 0)? " no_start_optimize" : "",
3104            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");
3105    
3106        if (jchanged) fprintf(outfile, "Duplicate name status changes\n");        if (jchanged) fprintf(outfile, "Duplicate name status changes\n");
# Line 1570  while (!done) Line 3141  while (!done)
3141          }          }
3142        else        else
3143          {          {
3144          int ch = first_char & 255;          const char *caseless =
3145          const char *caseless = ((first_char & REQ_CASELESS) == 0)?            ((((real_pcre *)re)->flags & PCRE_FCH_CASELESS) == 0)?
3146            "" : " (caseless)";            "" : " (caseless)";
3147          if (PRINTHEX(ch))  
3148            fprintf(outfile, "First char = \'%c\'%s\n", ch, caseless);          if (PRINTOK(first_char))
3149              fprintf(outfile, "First char = \'%c\'%s\n", first_char, caseless);
3150          else          else
3151            fprintf(outfile, "First char = %d%s\n", ch, caseless);            {
3152              fprintf(outfile, "First char = ");
3153              pchar(first_char, outfile);
3154              fprintf(outfile, "%s\n", caseless);
3155              }
3156          }          }
3157    
3158        if (need_char < 0)        if (need_char < 0)
# Line 1585  while (!done) Line 3161  while (!done)
3161          }          }
3162        else        else
3163          {          {
3164          int ch = need_char & 255;          const char *caseless =
3165          const char *caseless = ((need_char & REQ_CASELESS) == 0)?            ((((real_pcre *)re)->flags & PCRE_RCH_CASELESS) == 0)?
3166            "" : " (caseless)";            "" : " (caseless)";
3167          if (PRINTHEX(ch))  
3168            fprintf(outfile, "Need char = \'%c\'%s\n", ch, caseless);          if (PRINTOK(need_char))
3169              fprintf(outfile, "Need char = \'%c\'%s\n", need_char, caseless);
3170          else          else
3171            fprintf(outfile, "Need char = %d%s\n", ch, caseless);            {
3172              fprintf(outfile, "Need char = ");
3173              pchar(need_char, outfile);
3174              fprintf(outfile, "%s\n", caseless);
3175              }
3176          }          }
3177    
3178        /* Don't output study size; at present it is in any case a fixed        /* Don't output study size; at present it is in any case a fixed
3179        value, but it varies, depending on the computer architecture, and        value, but it varies, depending on the computer architecture, and
3180        so messes up the test suite. (And with the /F option, it might be        so messes up the test suite. (And with the /F option, it might be
3181        flipped.) */        flipped.) If study was forced by an external -s, don't show this
3182          information unless -i or -d was also present. This means that, except
3183          when auto-callouts are involved, the output from runs with and without
3184          -s should be identical. */
3185    
3186        if (do_study)        if (do_study || (force_study >= 0 && showinfo && !no_force_study))
3187          {          {
3188          if (extra == NULL)          if (extra == NULL)
3189            fprintf(outfile, "Study returned NULL\n");            fprintf(outfile, "Study returned NULL\n");
3190          else          else
3191            {            {
3192            uschar *start_bits = NULL;            pcre_uint8 *start_bits = NULL;
3193            new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits);            int minlength;
3194    
3195            if (start_bits == NULL)            if (new_info(re, extra, PCRE_INFO_MINLENGTH, &minlength) == 0)
3196              fprintf(outfile, "No starting byte set\n");              fprintf(outfile, "Subject length lower bound = %d\n", minlength);
3197            else  
3198              if (new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits) == 0)
3199              {              {
3200              int i;              if (start_bits == NULL)
3201              int c = 24;                fprintf(outfile, "No set of starting bytes\n");
3202              fprintf(outfile, "Starting byte set: ");              else
             for (i = 0; i < 256; i++)  
3203                {                {
3204                if ((start_bits[i/8] & (1<<(i&7))) != 0)                int i;
3205                  int c = 24;
3206                  fprintf(outfile, "Starting byte set: ");
3207                  for (i = 0; i < 256; i++)
3208                  {                  {
3209                  if (c > 75)                  if ((start_bits[i/8] & (1<<(i&7))) != 0)
3210                    {                    {
3211                    fprintf(outfile, "\n  ");                    if (c > 75)
3212                    c = 2;                      {
3213                    }                      fprintf(outfile, "\n  ");
3214                  if (PRINTHEX(i) && i != ' ')                      c = 2;
3215                    {                      }
3216                    fprintf(outfile, "%c ", i);                    if (PRINTOK(i) && i != ' ')
3217                    c += 2;                      {
3218                    }                      fprintf(outfile, "%c ", i);
3219                  else                      c += 2;
3220                    {                      }
3221                    fprintf(outfile, "\\x%02x ", i);                    else
3222                    c += 5;                      {
3223                        fprintf(outfile, "\\x%02x ", i);
3224                        c += 5;
3225                        }
3226                    }                    }
3227                  }                  }
3228                  fprintf(outfile, "\n");
3229                }                }
3230              fprintf(outfile, "\n");              }
3231              }
3232    
3233            /* Show this only if the JIT was set by /S, not by -s. */
3234    
3235            if ((study_options & PCRE_STUDY_JIT_COMPILE) != 0)
3236              {
3237              int jit;
3238              if (new_info(re, extra, PCRE_INFO_JIT, &jit) == 0)
3239                {
3240                if (jit)
3241                  fprintf(outfile, "JIT study was successful\n");
3242                else
3243    #ifdef SUPPORT_JIT
3244                  fprintf(outfile, "JIT study was not successful\n");
3245    #else
3246                  fprintf(outfile, "JIT support is not available in this version of PCRE\n");
3247    #endif
3248              }              }
3249            }            }
3250          }          }
# Line 1655  while (!done) Line 3263  while (!done)
3263          }          }
3264        else        else
3265          {          {
3266          uschar sbuf[8];          pcre_uint8 sbuf[8];
3267          sbuf[0] = (uschar)((true_size >> 24) & 255);  
3268          sbuf[1] = (uschar)((true_size >> 16) & 255);          if (do_flip) regexflip(re, extra);
3269          sbuf[2] = (uschar)((true_size >>  8) & 255);          sbuf[0] = (pcre_uint8)((true_size >> 24) & 255);
3270          sbuf[3] = (uschar)((true_size) & 255);          sbuf[1] = (pcre_uint8)((true_size >> 16) & 255);
3271            sbuf[2] = (pcre_uint8)((true_size >>  8) & 255);
3272          sbuf[4] = (uschar)((true_study_size >> 24) & 255);          sbuf[3] = (pcre_uint8)((true_size) & 255);
3273          sbuf[5] = (uschar)((true_study_size >> 16) & 255);          sbuf[4] = (pcre_uint8)((true_study_size >> 24) & 255);
3274          sbuf[6] = (uschar)((true_study_size >>  8) & 255);          sbuf[5] = (pcre_uint8)((true_study_size >> 16) & 255);
3275          sbuf[7] = (uschar)((true_study_size) & 255);          sbuf[6] = (pcre_uint8)((true_study_size >>  8) & 255);
3276            sbuf[7] = (pcre_uint8)((true_study_size) & 255);
3277    
3278          if (fwrite(sbuf, 1, 8, f) < 8 ||          if (fwrite(sbuf, 1, 8, f) < 8 ||
3279              fwrite(re, 1, true_size, f) < true_size)              fwrite(re, 1, true_size, f) < true_size)
# Line 1673  while (!done) Line 3282  while (!done)
3282            }            }
3283          else          else
3284            {            {
3285            fprintf(outfile, "Compiled regex written to %s\n", to_file);            fprintf(outfile, "Compiled pattern written to %s\n", to_file);
3286    
3287              /* If there is study data, write it. */
3288    
3289            if (extra != NULL)            if (extra != NULL)
3290              {              {
3291              if (fwrite(extra->study_data, 1, true_study_size, f) <              if (fwrite(extra->study_data, 1, true_study_size, f) <
# Line 1683  while (!done) Line 3295  while (!done)
3295                  strerror(errno));                  strerror(errno));
3296                }                }
3297              else fprintf(outfile, "Study data written to %s\n", to_file);              else fprintf(outfile, "Study data written to %s\n", to_file);
   
3298              }              }
3299            }            }
3300          fclose(f);          fclose(f);
3301          }          }
3302    
3303        new_free(re);        new_free(re);
3304        if (extra != NULL) new_free(extra);        if (extra != NULL)
3305        if (tables != NULL) new_free((void *)tables);          {
3306            PCRE_FREE_STUDY(extra);
3307            }
3308          if (locale_set)
3309            {
3310            new_free((void *)tables);
3311            setlocale(LC_CTYPE, "C");
3312            locale_set = 0;
3313            }
3314        continue;  /* With next regex */        continue;  /* With next regex */
3315        }        }
3316      }        /* End of non-POSIX compile */      }        /* End of non-POSIX compile */
# Line 1700  while (!done) Line 3319  while (!done)
3319    
3320    for (;;)    for (;;)
3321      {      {
3322      uschar *q;      pcre_uint8 *q;
3323      uschar *bptr;      pcre_uint8 *bptr;
3324      int *use_offsets = offsets;      int *use_offsets = offsets;
3325      int use_size_offsets = size_offsets;      int use_size_offsets = size_offsets;
3326      int callout_data = 0;      int callout_data = 0;
3327      int callout_data_set = 0;      int callout_data_set = 0;
3328      int count, c;      int count, c;
3329      int copystrings = 0;      int copystrings = 0;
3330      int find_match_limit = 0;      int find_match_limit = default_find_match_limit;
3331      int getstrings = 0;      int getstrings = 0;
3332      int getlist = 0;      int getlist = 0;
3333      int gmatched = 0;      int gmatched = 0;
3334      int start_offset = 0;      int start_offset = 0;
3335        int start_offset_sign = 1;
3336      int g_notempty = 0;      int g_notempty = 0;
3337      int use_dfa = 0;      int use_dfa = 0;
3338    
     options = 0;  
   
3339      *copynames = 0;      *copynames = 0;
3340      *getnames = 0;      *getnames = 0;
3341    
3342      copynamesptr = copynames;      cn16ptr = copynames;
3343      getnamesptr = getnames;      gn16ptr = getnames;
3344        cn8ptr = copynames8;
3345        gn8ptr = getnames8;
3346    
3347      pcre_callout = callout;      SET_PCRE_CALLOUT(callout);
3348      first_callout = 1;      first_callout = 1;
3349        last_callout_mark = NULL;
3350      callout_extra = 0;      callout_extra = 0;
3351      callout_count = 0;      callout_count = 0;
3352      callout_fail_count = 999999;      callout_fail_count = 999999;
3353      callout_fail_id = -1;      callout_fail_id = -1;
3354      show_malloc = 0;      show_malloc = 0;
3355        options = 0;
3356    
3357      if (extra != NULL) extra->flags &=      if (extra != NULL) extra->flags &=
3358        ~(PCRE_EXTRA_MATCH_LIMIT|PCRE_EXTRA_MATCH_LIMIT_RECURSION);        ~(PCRE_EXTRA_MATCH_LIMIT|PCRE_EXTRA_MATCH_LIMIT_RECURSION);
# Line 1740  while (!done) Line 3362  while (!done)
3362        {        {
3363        if (extend_inputline(infile, buffer + len, "data> ") == NULL)        if (extend_inputline(infile, buffer + len, "data> ") == NULL)
3364          {          {
3365          if (len > 0) break;          if (len > 0)    /* Reached EOF without hitting a newline */
3366              {
3367              fprintf(outfile, "\n");
3368              break;
3369              }
3370          done = 1;          done = 1;
3371          goto CONTINUE;          goto CONTINUE;
3372          }          }
# Line 1761  while (!done) Line 3387  while (!done)
3387        {        {
3388        int i = 0;        int i = 0;
3389        int n = 0;        int n = 0;
3390    
3391        if (c == '\\') switch ((c = *p++))        /* In UTF mode, input can be UTF-8, so just copy all non-backslash bytes.
3392          In non-UTF mode, allow the value of the byte to fall through to later,
3393          where values greater than 127 are turned into UTF-8 when running in
3394          16-bit mode. */
3395    
3396          if (c != '\\')
3397            {
3398            if (use_utf)
3399              {
3400              *q++ = c;
3401              continue;
3402              }
3403            }
3404    
3405          /* Handle backslash escapes */
3406    
3407          else switch ((c = *p++))
3408          {          {
3409          case 'a': c =    7; break;          case 'a': c =    7; break;
3410          case 'b': c = '\b'; break;          case 'b': c = '\b'; break;
# Line 1778  while (!done) Line 3420  while (!done)
3420          c -= '0';          c -= '0';
3421          while (i++ < 2 && isdigit(*p) && *p != '8' && *p != '9')          while (i++ < 2 && isdigit(*p) && *p != '8' && *p != '9')
3422            c = c * 8 + *p++ - '0';            c = c * 8 + *p++ - '0';
   
 #if !defined NOUTF8  
         if (use_utf8 && c > 255)  
           {  
           unsigned char buff8[8];  
           int ii, utn;  
           utn = ord2utf8(c, buff8);  
           for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];  
           c = buff8[ii];   /* Last byte */  
           }  
 #endif  
3423          break;          break;
3424    
3425          case 'x':          case 'x':
   
         /* Handle \x{..} specially - new Perl thing for utf8 */  
   
 #if !defined NOUTF8  
3426          if (*p == '{')          if (*p == '{')
3427            {            {
3428            unsigned char *pt = p;            pcre_uint8 *pt = p;
3429            c = 0;            c = 0;
3430            while (isxdigit(*(++pt)))  
3431              c = c * 16 + tolower(*pt) - ((isdigit(*pt))? '0' : 'W');            /* We used to have "while (isxdigit(*(++pt)))" here, but it fails
3432              when isxdigit() is a macro that refers to its argument more than
3433              once. This is banned by the C Standard, but apparently happens in at
3434              least one MacOS environment. */
3435    
3436              for (pt++; isxdigit(*pt); pt++)
3437                c = c * 16 + tolower(*pt) - ((isdigit(*pt))? '0' : 'a' - 10);
3438            if (*pt == '}')            if (*pt == '}')
3439              {              {
             unsigned char buff8[8];  
             int ii, utn;  
             if (use_utf8)  
               {  
               utn = ord2utf8(c, buff8);  
               for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];  
               c = buff8[ii];   /* Last byte */  
               }  
             else  
              {  
              if (c > 255)  
                fprintf(outfile, "** Character \\x{%x} is greater than 255 and "  
                  "UTF-8 mode is not enabled.\n"  
                  "** Truncation will probably give the wrong result.\n", c);  
              }  
3440              p = pt + 1;              p = pt + 1;
3441              break;              break;
3442              }              }
3443            /* Not correct form; fall through */            /* Not correct form for \x{...}; fall through */
3444            }            }
 #endif  
3445    
3446          /* Ordinary \x */          /* \x without {} always defines just one byte in 8-bit mode. This
3447            allows UTF-8 characters to be constructed byte by byte, and also allows
3448            invalid UTF-8 sequences to be made. Just copy the byte in UTF mode.
3449            Otherwise, pass it down to later code so that it can be turned into
3450            UTF-8 when running in 16-bit mode. */
3451    
3452          c = 0;          c = 0;
3453          while (i++ < 2 && isxdigit(*p))          while (i++ < 2 && isxdigit(*p))
3454            {            {
3455            c = c * 16 + tolower(*p) - ((isdigit(*p))? '0' : 'W');            c = c * 16 + tolower(*p) - ((isdigit(*p))? '0' : 'a' - 10);
3456            p++;            p++;
3457            }            }
3458            if (use_utf)
3459              {
3460              *q++ = c;
3461              continue;
3462              }
3463          break;          break;
3464    
3465          case 0:   /* \ followed by EOF allows for an empty line */          case 0:   /* \ followed by EOF allows for an empty line */
# Line 1841  while (!done) Line 3467  while (!done)
3467          continue;          continue;
3468    
3469          case '>':          case '>':
3470            if (*p == '-')
3471              {
3472              start_offset_sign = -1;
3473              p++;
3474              }
3475          while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0';          while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0';
3476            start_offset *= start_offset_sign;
3477          continue;          continue;
3478    
3479          case 'A':  /* Option setting */          case 'A':  /* Option setting */
# Line 1860  while (!done) Line 3492  while (!done)
3492            }            }
3493          else if (isalnum(*p))          else if (isalnum(*p))
3494            {            {
3495            uschar *npp = copynamesptr;            READ_CAPTURE_NAME(p, &cn8ptr, &cn16ptr, re);
           while (isalnum(*p)) *npp++ = *p++;  
           *npp++ = 0;  
           *npp = 0;  
           n = pcre_get_stringnumber(re, (char *)copynamesptr);  
           if (n < 0)  
             fprintf(outfile, "no parentheses with name \"%s\"\n", copynamesptr);  
           copynamesptr = npp;  
3496            }            }
3497          else if (*p == '+')          else if (*p == '+')
3498            {            {
# Line 1876  while (!done) Line 3501  while (!done)
3501            }            }
3502          else if (*p == '-')          else if (*p == '-')
3503            {            {
3504            pcre_callout = NULL;            SET_PCRE_CALLOUT(NULL);
3505            p++;            p++;
3506            }            }
3507          else if (*p == '!')          else if (*p == '!')
# Line 1914  while (!done) Line 3539  while (!done)
3539  #endif  #endif
3540            use_dfa = 1;            use_dfa = 1;
3541          continue;          continue;
3542    #endif
3543    
3544    #if !defined NODFA
3545          case 'F':          case 'F':
3546          options |= PCRE_DFA_SHORTEST;          options |= PCRE_DFA_SHORTEST;
3547          continue;          continue;
# Line 1928  while (!done) Line 3555  while (!done)
3555            }            }
3556          else if (isalnum(*p))          else if (isalnum(*p))
3557            {            {
3558            uschar *npp = getnamesptr;            READ_CAPTURE_NAME(p, &gn8ptr, &gn16ptr, re);
3559            while (isalnum(*p)) *npp++ = *p++;            }
3560            *npp++ = 0;          continue;
3561            *npp = 0;  
3562            n = pcre_get_stringnumber(re, (char *)getnamesptr);          case 'J':
3563            if (n < 0)          while(isdigit(*p)) n = n * 10 + *p++ - '0';
3564              fprintf(outfile, "no parentheses with name \"%s\"\n", getnamesptr);          if (extra != NULL
3565            getnamesptr = npp;              && (extra->flags & PCRE_EXTRA_EXECUTABLE_JIT) != 0
3566                && extra->executable_jit != NULL)
3567              {
3568              if (jit_stack != NULL) PCRE_JIT_STACK_FREE(jit_stack);
3569              jit_stack = PCRE_JIT_STACK_ALLOC(1, n * 1024);
3570              PCRE_ASSIGN_JIT_STACK(extra, jit_callback, jit_stack);
3571            }            }
3572          continue;          continue;
3573    
# Line 1948  while (!done) Line 3580  while (!done)
3580          continue;          continue;
3581    
3582          case 'N':          case 'N':
3583          options |= PCRE_NOTEMPTY;          if ((options & PCRE_NOTEMPTY) != 0)
3584              options = (options & ~PCRE_NOTEMPTY) | PCRE_NOTEMPTY_ATSTART;
3585            else
3586              options |= PCRE_NOTEMPTY;
3587          continue;          continue;
3588    
3589          case 'O':          case 'O':
# Line 1971  while (!done) Line 3606  while (!done)
3606          continue;          continue;
3607    
3608          case 'P':          case 'P':
3609          options |= PCRE_PARTIAL;          options |= ((options & PCRE_PARTIAL_SOFT) == 0)?
3610              PCRE_PARTIAL_SOFT : PCRE_PARTIAL_HARD;
3611          continue;          continue;
3612    
3613          case 'Q':          case 'Q':
# Line 2006  while (!done) Line 3642  while (!done)
3642          show_malloc = 1;          show_malloc = 1;
3643          continue;          continue;
3644    
3645            case 'Y':
3646            options |= PCRE_NO_START_OPTIMIZE;
3647            continue;
3648    
3649          case 'Z':          case 'Z':
3650          options |= PCRE_NOTEOL;          options |= PCRE_NOTEOL;
3651          continue;          continue;
# Line 2023  while (!done) Line 3663  while (!done)
3663            }            }
3664          continue;          continue;
3665          }          }
3666        *q++ = c;  
3667          /* We now have a character value in c that may be greater than 255. In
3668          16-bit mode, we always convert characters to UTF-8 so that values greater
3669          than 255 can be passed to non-UTF 16-bit strings. In 8-bit mode we
3670          convert to UTF-8 if we are in UTF mode. Values greater than 127 in UTF
3671          mode must have come from \x{...} or octal constructs because values from
3672          \x.. get this far only in non-UTF mode. */
3673    
3674          if (use_pcre16 || use_utf)
3675            {
3676            pcre_uint8 buff8[8];
3677            int ii, utn;
3678            utn = ord2utf8(c, buff8);
3679            for (ii = 0; ii < utn; ii++) *q++ = buff8[ii];
3680            }
3681          else
3682            {
3683            if (c > 255)
3684              {
3685              fprintf(outfile, "** Character \\x{%x} is greater than 255 "
3686                "and UTF-8 mode is not enabled.\n", c);
3687              fprintf(outfile, "** Truncation will probably give the wrong "
3688                "result.\n");
3689              }
3690            *q++ = c;
3691            }
3692        }        }
3693    
3694        /* Reached end of subject string */
3695    
3696      *q = 0;      *q = 0;
3697      len = q - dbuffer;      len = (int)(q - dbuffer);
3698    
3699        /* Move the data to the end of the buffer so that a read over the end of
3700        the buffer will be seen by valgrind, even if it doesn't cause a crash. If
3701        we are using the POSIX interface, we must include the terminating zero. */
3702    
3703    #if !defined NOPOSIX
3704        if (posix || do_posix)
3705          {
3706          memmove(bptr + buffer_size - len - 1, bptr, len + 1);
3707          bptr += buffer_size - len - 1;
3708          }
3709        else
3710    #endif
3711          {
3712          memmove(bptr + buffer_size - len, bptr, len);
3713          bptr += buffer_size - len;
3714          }
3715    
3716      if ((all_use_dfa || use_dfa) && find_match_limit)      if ((all_use_dfa || use_dfa) && find_match_limit)
3717        {        {
# Line 2047  while (!done) Line 3732  while (!done)
3732          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);
3733        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;
3734        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;
3735          if ((options & PCRE_NOTEMPTY) != 0) eflags |= REG_NOTEMPTY;
3736    
3737        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);
3738    
# Line 2068  while (!done) Line 3754  while (!done)
3754            if (pmatch[i].rm_so >= 0)            if (pmatch[i].rm_so >= 0)
3755              {              {
3756              fprintf(outfile, "%2d: ", (int)i);              fprintf(outfile, "%2d: ", (int)i);
3757              (void)pchars(dbuffer + pmatch[i].rm_so,              PCHARSV(dbuffer, pmatch[i].rm_so,
3758                pmatch[i].rm_eo - pmatch[i].rm_so, outfile);                pmatch[i].rm_eo - pmatch[i].rm_so, outfile);
3759              fprintf(outfile, "\n");              fprintf(outfile, "\n");
3760              if (i == 0 && do_showrest)              if (do_showcaprest || (i == 0 && do_showrest))
3761                {                {
3762                fprintf(outfile, " 0+ ");                fprintf(outfile, "%2d+ ", (int)i);
3763                (void)pchars(dbuffer + pmatch[i].rm_eo, len - pmatch[i].rm_eo,                PCHARSV(dbuffer, pmatch[i].rm_eo, len - pmatch[i].rm_eo,
3764                  outfile);                  outfile);
3765                fprintf(outfile, "\n");                fprintf(outfile, "\n");
3766                }                }
# Line 2082  while (!done) Line 3768  while (!done)
3768            }            }
3769          }          }
3770        free(pmatch);        free(pmatch);
3771          goto NEXT_DATA;
3772        }        }
3773    
3774    #endif  /* !defined NOPOSIX */
3775    
3776      /* Handle matching via the native interface - repeats for /g and /G */      /* Handle matching via the native interface - repeats for /g and /G */
3777    
3778      else  #ifdef SUPPORT_PCRE16
3779  #endif  /* !defined NOPOSIX */      if (use_pcre16)
3780          {
3781          len = to16(TRUE, bptr, (((real_pcre *)re)->options) & PCRE_UTF8, len);
3782          switch(len)
3783            {
3784            case -1:
3785            fprintf(outfile, "**Failed: invalid UTF-8 string cannot be "
3786              "converted to UTF-16\n");
3787            goto NEXT_DATA;
3788    
3789            case -2:
3790            fprintf(outfile, "**Failed: character value greater than 0x10ffff "
3791              "cannot be converted to UTF-16\n");
3792            goto NEXT_DATA;
3793    
3794            case -3:
3795            fprintf(outfile, "**Failed: character value greater than 0xffff "
3796              "cannot be converted to 16-bit in non-UTF mode\n");
3797            goto NEXT_DATA;
3798    
3799            default:
3800            break;
3801            }
3802          bptr = (pcre_uint8 *)buffer16;
3803          }
3804    #endif
3805    
3806      for (;; gmatched++)    /* Loop for /g or /G */      for (;; gmatched++)    /* Loop for /g or /G */
3807        {        {
3808          markptr = NULL;
3809    
3810        if (timeitm > 0)        if (timeitm > 0)
3811          {          {
3812          register int i;          register int i;
# Line 2102  while (!done) Line 3818  while (!done)
3818            {            {
3819            int workspace[1000];            int workspace[1000];
3820            for (i = 0; i < timeitm; i++)            for (i = 0; i < timeitm; i++)
3821              count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,              {
3822                options | g_notempty, use_offsets, use_size_offsets, workspace,              PCRE_DFA_EXEC(count, re, extra, bptr, len, start_offset,
3823                sizeof(workspace)/sizeof(int));                (options | g_notempty), use_offsets, use_size_offsets, workspace,
3824                  (sizeof(workspace)/sizeof(int)));
3825                }
3826            }            }
3827          else          else
3828  #endif  #endif
3829    
3830          for (i = 0; i < timeitm; i++)          for (i = 0; i < timeitm; i++)
3831            count = pcre_exec(re, extra, (char *)bptr, len,            {
3832              start_offset, options | g_notempty, use_offsets, use_size_offsets);            PCRE_EXEC(count, re, extra, bptr, len, start_offset,
3833                (options | g_notempty), use_offsets, use_size_offsets);
3834              }
3835          time_taken = clock() - start_time;          time_taken = clock() - start_time;
3836          fprintf(outfile, "Execute time %.4f milliseconds\n",          fprintf(outfile, "Execute time %.4f milliseconds\n",
3837            (((double)time_taken * 1000.0) / (double)timeitm) /            (((double)time_taken * 1000.0) / (double)timeitm) /
# Line 2121  while (!done) Line 3840  while (!done)
3840    
3841        /* If find_match_limit is set, we want to do repeated matches with        /* If find_match_limit is set, we want to do repeated matches with
3842        varying limits in order to find the minimum value for the match limit and        varying limits in order to find the minimum value for the match limit and
3843        for the recursion limit. */        for the recursion limit. The match limits are relevant only to the normal
3844          running of pcre_exec(), so disable the JIT optimization. This makes it
3845          possible to run the same set of tests with and without JIT externally
3846          requested. */
3847    
3848        if (find_match_limit)        if (find_match_limit)
3849          {          {
# Line 2130  while (!done) Line 3852  while (!done)
3852            extra = (pcre_extra *)malloc(sizeof(pcre_extra));            extra = (pcre_extra *)malloc(sizeof(pcre_extra));
3853            extra->flags = 0;            extra->flags = 0;
3854            }            }
3855            else extra->flags &= ~PCRE_EXTRA_EXECUTABLE_JIT;
3856    
3857          (void)check_match_limit(re, extra, bptr, len, start_offset,          (void)check_match_limit(re, extra, bptr, len, start_offset,
3858            options|g_notempty, use_offsets, use_size_offsets,            options|g_notempty, use_offsets, use_size_offsets,
# Line 2153  while (!done) Line 3876  while (!done)
3876            }            }
3877          extra->flags |= PCRE_EXTRA_CALLOUT_DATA;          extra->flags |= PCRE_EXTRA_CALLOUT_DATA;
3878          extra->callout_data = &callout_data;          extra->callout_data = &callout_data;
3879          count = pcre_exec(re, extra, (char *)bptr, len, start_offset,          PCRE_EXEC(count, re, extra, bptr, len, start_offset,
3880            options | g_notempty, use_offsets, use_size_offsets);            options | g_notempty, use_offsets, use_size_offsets);
3881          extra->flags &= ~PCRE_EXTRA_CALLOUT_DATA;          extra->flags &= ~PCRE_EXTRA_CALLOUT_DATA;
3882          }          }
# Line 2165  while (!done) Line 3888  while (!done)
3888        else if (all_use_dfa || use_dfa)        else if (all_use_dfa || use_dfa)
3889          {          {
3890          int workspace[1000];          int workspace[1000];
3891          count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,          PCRE_DFA_EXEC(count, re, extra, bptr, len, start_offset,
3892            options | g_notempty, use_offsets, use_size_offsets, workspace,            (options | g_notempty), use_offsets, use_size_offsets, workspace,
3893            sizeof(workspace)/sizeof(int));            (sizeof(workspace)/sizeof(int)));
3894          if (count == 0)          if (count == 0)
3895            {            {
3896            fprintf(outfile, "Matched, but too many subsidiary matches\n");            fprintf(outfile, "Matched, but too many subsidiary matches\n");
# Line 2178  while (!done) Line 3901  while (!done)
3901    
3902        else        else
3903          {          {
3904          count = pcre_exec(re, extra, (char *)bptr, len,          PCRE_EXEC(count, re, extra, bptr, len, start_offset,
3905            start_offset, options | g_notempty, use_offsets, use_size_offsets);            options | g_notempty, use_offsets, use_size_offsets);
3906          if (count == 0)          if (count == 0)
3907            {            {
3908            fprintf(outfile, "Matched, but too many substrings\n");            fprintf(outfile, "Matched, but too many substrings\n");
# Line 2192  while (!done) Line 3915  while (!done)
3915        if (count >= 0)        if (count >= 0)
3916          {          {
3917          int i, maxcount;          int i, maxcount;
3918            void *cnptr, *gnptr;
3919    
3920  #if !defined NODFA  #if !defined NODFA
3921          if (all_use_dfa || use_dfa) maxcount = use_size_offsets/2; else          if (all_use_dfa || use_dfa) maxcount = use_size_offsets/2; else
# Line 2213  while (!done) Line 3937  while (!done)
3937              }              }
3938            }            }
3939    
3940            /* do_allcaps requests showing of all captures in the pattern, to check
3941            unset ones at the end. */
3942    
3943            if (do_allcaps)
3944              {
3945              if (new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count) < 0)
3946                goto SKIP_DATA;
3947              count++;   /* Allow for full match */
3948              if (count * 2 > use_size_offsets) count = use_size_offsets/2;
3949              }
3950    
3951            /* Output the captured substrings */
3952    
3953          for (i = 0; i < count * 2; i += 2)          for (i = 0; i < count * 2; i += 2)
3954            {            {
3955            if (use_offsets[i] < 0)            if (use_offsets[i] < 0)
3956                {
3957                if (use_offsets[i] != -1)
3958                  fprintf(outfile, "ERROR: bad negative value %d for offset %d\n",
3959                    use_offsets[i], i);
3960                if (use_offsets[i+1] != -1)
3961                  fprintf(outfile, "ERROR: bad negative value %d for offset %d\n",
3962                    use_offsets[i+1], i+1);
3963              fprintf(outfile, "%2d: <unset>\n", i/2);              fprintf(outfile, "%2d: <unset>\n", i/2);
3964                }
3965            else            else
3966              {              {
3967              fprintf(outfile, "%2d: ", i/2);              fprintf(outfile, "%2d: ", i/2);
3968              (void)pchars(bptr + use_offsets[i],              PCHARSV(bptr, use_offsets[i],
3969                use_offsets[i+1] - use_offsets[i], outfile);                use_offsets[i+1] - use_offsets[i], outfile);
3970              fprintf(outfile, "\n");              fprintf(outfile, "\n");
3971              if (i == 0)              if (do_showcaprest || (i == 0 && do_showrest))
3972                {                {
3973                if (do_showrest)                fprintf(outfile, "%2d+ ", i/2);
3974                  {                PCHARSV(bptr, use_offsets[i+1], len - use_offsets[i+1],
3975                  fprintf(outfile, " 0+ ");                  outfile);
3976                  (void)pchars(bptr + use_offsets[i+1], len - use_offsets[i+1],                fprintf(outfile, "\n");
                   outfile);  
                 fprintf(outfile, "\n");  
                 }  
3977                }                }
3978              }              }
3979            }            }
3980    
3981            if (markptr != NULL)
3982              {
3983              fprintf(outfile, "MK: ");
3984              PCHARSV(markptr, 0, -1, outfile);
3985              fprintf(outfile, "\n");
3986              }
3987    
3988          for (i = 0; i < 32; i++)          for (i = 0; i < 32; i++)
3989            {            {
3990            if ((copystrings & (1 << i)) != 0)            if ((copystrings & (1 << i)) != 0)
3991              {              {
3992                int rc;
3993              char copybuffer[256];              char copybuffer[256];
3994              int rc = pcre_copy_substring((char *)bptr, use_offsets, count,              PCRE_COPY_SUBSTRING(rc, bptr, use_offsets, count, i,
3995                i, copybuffer, sizeof(copybuffer));                copybuffer, sizeof(copybuffer));
3996              if (rc < 0)              if (rc < 0)
3997                fprintf(outfile, "copy substring %d failed %d\n", i, rc);                fprintf(outfile, "copy substring %d failed %d\n", i, rc);
3998              else              else
3999                fprintf(outfile, "%2dC %s (%d)\n", i, copybuffer, rc);                {
4000                  fprintf(outfile, "%2dC ", i);
4001                  PCHARSV(copybuffer, 0, rc, outfile);
4002                  fprintf(outfile, " (%d)\n", rc);
4003                  }
4004              }              }
4005            }            }
4006    
4007          for (copynamesptr = copynames;          cnptr = copynames;
4008               *copynamesptr != 0;          for (;;)
              copynamesptr += (int)strlen((char*)copynamesptr) + 1)  
4009            {            {
4010              int rc;
4011            char copybuffer[256];            char copybuffer[256];
4012            int rc = pcre_copy_named_substring(re, (char *)bptr, use_offsets,  
4013              count, (char *)copynamesptr, copybuffer, sizeof(copybuffer));            if (use_pcre16)
4014                {
4015                if (*(pcre_uint16 *)cnptr == 0) break;
4016                }
4017              else
4018                {
4019                if (*(pcre_uint8 *)cnptr == 0) break;
4020                }
4021    
4022              PCRE_COPY_NAMED_SUBSTRING(rc, re, bptr, use_offsets, count,
4023                cnptr, copybuffer, sizeof(copybuffer));
4024    
4025            if (rc < 0)            if (rc < 0)
4026              fprintf(outfile, "copy substring %s failed %d\n", copynamesptr, rc);              {
4027                fprintf(outfile, "copy substring ");
4028                PCHARSV(cnptr, 0, -1, outfile);
4029                fprintf(outfile, " failed %d\n", rc);
4030                }
4031            else            else
4032              fprintf(outfile, "  C %s (%d) %s\n", copybuffer, rc, copynamesptr);              {
4033                fprintf(outfile, "  C ");
4034                PCHARSV(copybuffer, 0, rc, outfile);
4035                fprintf(outfile, " (%d) ", rc);
4036                PCHARSV(cnptr, 0, -1, outfile);
4037                putc('\n', outfile);
4038                }
4039    
4040              cnptr = (char *)cnptr + (STRLEN(cnptr) + 1) * CHAR_SIZE;
4041            }            }
4042    
4043          for (i = 0; i < 32; i++)          for (i = 0; i < 32; i++)
4044            {            {
4045            if ((getstrings & (1 << i)) != 0)            if ((getstrings & (1 << i)) != 0)
4046              {              {
4047                int rc;
4048              const char *substring;              const char *substring;
4049              int rc = pcre_get_substring((char *)bptr, use_offsets, count,              PCRE_GET_SUBSTRING(rc, bptr, use_offsets, count, i, &substring);
               i, &substring);  
4050              if (rc < 0)              if (rc < 0)
4051                fprintf(outfile, "get substring %d failed %d\n", i, rc);                fprintf(outfile, "get substring %d failed %d\n", i, rc);
4052              else              else
4053                {                {
4054                fprintf(outfile, "%2dG %s (%d)\n", i, substring, rc);                fprintf(outfile, "%2dG ", i);
4055                pcre_free_substring(substring);                PCHARSV(substring, 0, rc, outfile);
4056                  fprintf(outfile, " (%d)\n", rc);
4057                  PCRE_FREE_SUBSTRING(substring);
4058                }                }
4059              }              }
4060            }            }
4061    
4062          for (getnamesptr = getnames;          gnptr = getnames;
4063               *getnamesptr != 0;          for (;;)
              getnamesptr += (int)strlen((char*)getnamesptr) + 1)  
4064            {            {
4065              int rc;
4066            const char *substring;            const char *substring;
4067            int rc = pcre_get_named_substring(re, (char *)bptr, use_offsets,  
4068              count, (char *)getnamesptr, &substring);            if (use_pcre16)
4069                {
4070                if (*(pcre_uint16 *)gnptr == 0) break;
4071                }
4072              else
4073                {
4074                if (*(pcre_uint8 *)gnptr == 0) break;
4075                }
4076    
4077              PCRE_GET_NAMED_SUBSTRING(rc, re, bptr, use_offsets, count,
4078                gnptr, &substring);
4079            if (rc < 0)            if (rc < 0)
4080              fprintf(outfile, "copy substring %s failed %d\n", getnamesptr, rc);              {
4081                fprintf(outfile, "get substring ");
4082                PCHARSV(gnptr, 0, -1, outfile);
4083                fprintf(outfile, " failed %d\n", rc);
4084                }
4085            else            else
4086              {              {
4087              fprintf(outfile, "  G %s (%d) %s\n", substring, rc, getnamesptr);              fprintf(outfile, "  G ");
4088              pcre_free_substring(substring);              PCHARSV(substring, 0, rc, outfile);
4089                fprintf(outfile, " (%d) ", rc);
4090                PCHARSV(gnptr, 0, -1, outfile);
4091                PCRE_FREE_SUBSTRING(substring);
4092                putc('\n', outfile);
4093              }              }
4094    
4095              gnptr = (char *)gnptr + (STRLEN(gnptr) + 1) * CHAR_SIZE;
4096            }            }
4097    
4098          if (getlist)          if (getlist)
4099            {            {
4100              int rc;
4101            const char **stringlist;            const char **stringlist;
4102            int rc = pcre_get_substring_list((char *)bptr, use_offsets, count,            PCRE_GET_SUBSTRING_LIST(rc, bptr, use_offsets, count, &stringlist);
             &stringlist);  
4103            if (rc < 0)            if (rc < 0)
4104              fprintf(outfile, "get substring list failed %d\n", rc);              fprintf(outfile, "get substring list failed %d\n", rc);
4105            else            else
4106              {              {
4107              for (i = 0; i < count; i++)              for (i = 0; i < count; i++)
4108                fprintf(outfile, "%2dL %s\n", i, stringlist[i]);                {
4109                  fprintf(outfile, "%2dL ", i);
4110                  PCHARSV(stringlist[i], 0, -1, outfile);
4111                  putc('\n', outfile);
4112                  }
4113              if (stringlist[i] != NULL)              if (stringlist[i] != NULL)
4114                fprintf(outfile, "string list not terminated by NULL\n");                fprintf(outfile, "string list not terminated by NULL\n");
4115              /* free((void *)stringlist); */              PCRE_FREE_SUBSTRING_LIST(stringlist);
             pcre_free_substring_list(stringlist);  
4116              }              }
4117            }            }
4118          }          }
# Line 2319  while (!done) Line 4121  while (!done)
4121    
4122        else if (count == PCRE_ERROR_PARTIAL)        else if (count == PCRE_ERROR_PARTIAL)
4123          {          {
4124